COMPMID-3639: (3RDPARTY_UPDATE) Move CL kernels to src

Change-Id: I10d27db788e5086adae1841e3e2441cd9b76ef84
Signed-off-by: Sang-Hoon Park <sang-hoon.park@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4310
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>

140 files changed, 35 insertions, 14265 deletions

diff --git a/arm_compute/core/CL/CLKernels.h b/arm_compute/core/CL/CLKernels.h
deleted file mode 100644
index 231534fe50..0000000000
--- a/arm_compute/core/CL/CLKernels.h
+++ /dev/null
@@ -1,163 +0,0 @@
-/*
- * Copyright (c) 2016-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLKERNELS_H
-#define ARM_COMPUTE_CLKERNELS_H
-
-/* Header regrouping all the CL kernels */
-#include "arm_compute/core/CL/kernels/CLAbsoluteDifferenceKernel.h"
-#include "arm_compute/core/CL/kernels/CLAccumulateKernel.h"
-#include "arm_compute/core/CL/kernels/CLActivationLayerKernel.h"
-#include "arm_compute/core/CL/kernels/CLArgMinMaxLayerKernel.h"
-#include "arm_compute/core/CL/kernels/CLBatchConcatenateLayerKernel.h"
-#include "arm_compute/core/CL/kernels/CLBatchNormalizationLayerKernel.h"
-#include "arm_compute/core/CL/kernels/CLBatchToSpaceLayerKernel.h"
-#include "arm_compute/core/CL/kernels/CLBitwiseAndKernel.h"
-#include "arm_compute/core/CL/kernels/CLBitwiseNotKernel.h"
-#include "arm_compute/core/CL/kernels/CLBitwiseOrKernel.h"
-#include "arm_compute/core/CL/kernels/CLBitwiseXorKernel.h"
-#include "arm_compute/core/CL/kernels/CLBoundingBoxTransformKernel.h"
-#include "arm_compute/core/CL/kernels/CLBox3x3Kernel.h"
-#include "arm_compute/core/CL/kernels/CLCannyEdgeKernel.h"
-#include "arm_compute/core/CL/kernels/CLChannelCombineKernel.h"
-#include "arm_compute/core/CL/kernels/CLChannelExtractKernel.h"
-#include "arm_compute/core/CL/kernels/CLChannelShuffleLayerKernel.h"
-#include "arm_compute/core/CL/kernels/CLCol2ImKernel.h"
-#include "arm_compute/core/CL/kernels/CLColorConvertKernel.h"
-#include "arm_compute/core/CL/kernels/CLComparisonKernel.h"
-#include "arm_compute/core/CL/kernels/CLConvertFullyConnectedWeightsKernel.h"
-#include "arm_compute/core/CL/kernels/CLConvolutionKernel.h"
-#include "arm_compute/core/CL/kernels/CLCopyKernel.h"
-#include "arm_compute/core/CL/kernels/CLCropKernel.h"
-#include "arm_compute/core/CL/kernels/CLDeconvolutionLayerUpsampleKernel.h"
-#include "arm_compute/core/CL/kernels/CLDeconvolutionReshapeOutputKernel.h"
-#include "arm_compute/core/CL/kernels/CLDepthConcatenateLayerKernel.h"
-#include "arm_compute/core/CL/kernels/CLDepthConvertLayerKernel.h"
-#include "arm_compute/core/CL/kernels/CLDepthToSpaceLayerKernel.h"
-#include "arm_compute/core/CL/kernels/CLDepthwiseConvolutionLayer3x3NCHWKernel.h"
-#include "arm_compute/core/CL/kernels/CLDepthwiseConvolutionLayer3x3NHWCKernel.h"
-#include "arm_compute/core/CL/kernels/CLDepthwiseConvolutionLayerNativeKernel.h"
-#include "arm_compute/core/CL/kernels/CLDepthwiseConvolutionLayerReshapeWeightsKernel.h"
-#include "arm_compute/core/CL/kernels/CLDequantizationLayerKernel.h"
-#include "arm_compute/core/CL/kernels/CLDerivativeKernel.h"
-#include "arm_compute/core/CL/kernels/CLDilateKernel.h"
-#include "arm_compute/core/CL/kernels/CLDirectConvolutionLayerKernel.h"
-#include "arm_compute/core/CL/kernels/CLElementWiseUnaryLayerKernel.h"
-#include "arm_compute/core/CL/kernels/CLElementwiseOperationKernel.h"
-#include "arm_compute/core/CL/kernels/CLErodeKernel.h"
-#include "arm_compute/core/CL/kernels/CLFFTDigitReverseKernel.h"
-#include "arm_compute/core/CL/kernels/CLFFTRadixStageKernel.h"
-#include "arm_compute/core/CL/kernels/CLFFTScaleKernel.h"
-#include "arm_compute/core/CL/kernels/CLFastCornersKernel.h"
-#include "arm_compute/core/CL/kernels/CLFillBorderKernel.h"
-#include "arm_compute/core/CL/kernels/CLFlattenLayerKernel.h"
-#include "arm_compute/core/CL/kernels/CLFloorKernel.h"
-#include "arm_compute/core/CL/kernels/CLFuseBatchNormalizationKernel.h"
-#include "arm_compute/core/CL/kernels/CLGEMMLowpMatrixMultiplyNativeKernel.h"
-#include "arm_compute/core/CL/kernels/CLGEMMLowpMatrixMultiplyReshapedKernel.h"
-#include "arm_compute/core/CL/kernels/CLGEMMLowpMatrixMultiplyReshapedOnlyRHSKernel.h"
-#include "arm_compute/core/CL/kernels/CLGEMMLowpOffsetContributionKernel.h"
-#include "arm_compute/core/CL/kernels/CLGEMMLowpOffsetContributionOutputStageKernel.h"
-#include "arm_compute/core/CL/kernels/CLGEMMLowpQuantizeDownInt32ScaleByFixedPointKernel.h"
-#include "arm_compute/core/CL/kernels/CLGEMMLowpQuantizeDownInt32ScaleByFloatKernel.h"
-#include "arm_compute/core/CL/kernels/CLGEMMLowpQuantizeDownInt32ScaleKernel.h"
-#include "arm_compute/core/CL/kernels/CLGEMMLowpReductionKernel.h"
-#include "arm_compute/core/CL/kernels/CLGEMMMatrixMultiplyKernel.h"
-#include "arm_compute/core/CL/kernels/CLGEMMMatrixMultiplyNativeKernel.h"
-#include "arm_compute/core/CL/kernels/CLGEMMMatrixMultiplyReshapedKernel.h"
-#include "arm_compute/core/CL/kernels/CLGEMMMatrixMultiplyReshapedOnlyRHSKernel.h"
-#include "arm_compute/core/CL/kernels/CLGEMMMatrixVectorMultiplyKernel.h"
-#include "arm_compute/core/CL/kernels/CLGEMMReshapeLHSMatrixKernel.h"
-#include "arm_compute/core/CL/kernels/CLGEMMReshapeRHSMatrixKernel.h"
-#include "arm_compute/core/CL/kernels/CLGatherKernel.h"
-#include "arm_compute/core/CL/kernels/CLGaussian3x3Kernel.h"
-#include "arm_compute/core/CL/kernels/CLGaussian5x5Kernel.h"
-#include "arm_compute/core/CL/kernels/CLGaussianPyramidKernel.h"
-#include "arm_compute/core/CL/kernels/CLGenerateProposalsLayerKernel.h"
-#include "arm_compute/core/CL/kernels/CLHOGDescriptorKernel.h"
-#include "arm_compute/core/CL/kernels/CLHOGDetectorKernel.h"
-#include "arm_compute/core/CL/kernels/CLHarrisCornersKernel.h"
-#include "arm_compute/core/CL/kernels/CLHeightConcatenateLayerKernel.h"
-#include "arm_compute/core/CL/kernels/CLHistogramKernel.h"
-#include "arm_compute/core/CL/kernels/CLIm2ColKernel.h"
-#include "arm_compute/core/CL/kernels/CLInstanceNormalizationLayerKernel.h"
-#include "arm_compute/core/CL/kernels/CLIntegralImageKernel.h"
-#include "arm_compute/core/CL/kernels/CLL2NormalizeLayerKernel.h"
-#include "arm_compute/core/CL/kernels/CLLKTrackerKernel.h"
-#include "arm_compute/core/CL/kernels/CLLocallyConnectedMatrixMultiplyKernel.h"
-#include "arm_compute/core/CL/kernels/CLMagnitudePhaseKernel.h"
-#include "arm_compute/core/CL/kernels/CLMaxUnpoolingLayerKernel.h"
-#include "arm_compute/core/CL/kernels/CLMeanStdDevKernel.h"
-#include "arm_compute/core/CL/kernels/CLMeanStdDevNormalizationKernel.h"
-#include "arm_compute/core/CL/kernels/CLMedian3x3Kernel.h"
-#include "arm_compute/core/CL/kernels/CLMemsetKernel.h"
-#include "arm_compute/core/CL/kernels/CLMinMaxLayerKernel.h"
-#include "arm_compute/core/CL/kernels/CLMinMaxLocationKernel.h"
-#include "arm_compute/core/CL/kernels/CLNonLinearFilterKernel.h"
-#include "arm_compute/core/CL/kernels/CLNonMaximaSuppression3x3Kernel.h"
-#include "arm_compute/core/CL/kernels/CLNormalizationLayerKernel.h"
-#include "arm_compute/core/CL/kernels/CLNormalizePlanarYUVLayerKernel.h"
-#include "arm_compute/core/CL/kernels/CLPadLayerKernel.h"
-#include "arm_compute/core/CL/kernels/CLPermuteKernel.h"
-#include "arm_compute/core/CL/kernels/CLPixelWiseMultiplicationKernel.h"
-#include "arm_compute/core/CL/kernels/CLPoolingLayerKernel.h"
-#include "arm_compute/core/CL/kernels/CLPriorBoxLayerKernel.h"
-#include "arm_compute/core/CL/kernels/CLQLSTMLayerNormalizationKernel.h"
-#include "arm_compute/core/CL/kernels/CLQuantizationLayerKernel.h"
-#include "arm_compute/core/CL/kernels/CLROIAlignLayerKernel.h"
-#include "arm_compute/core/CL/kernels/CLROIPoolingLayerKernel.h"
-#include "arm_compute/core/CL/kernels/CLRangeKernel.h"
-#include "arm_compute/core/CL/kernels/CLReductionOperationKernel.h"
-#include "arm_compute/core/CL/kernels/CLRemapKernel.h"
-#include "arm_compute/core/CL/kernels/CLReorgLayerKernel.h"
-#include "arm_compute/core/CL/kernels/CLReshapeLayerKernel.h"
-#include "arm_compute/core/CL/kernels/CLReverseKernel.h"
-#include "arm_compute/core/CL/kernels/CLScaleKernel.h"
-#include "arm_compute/core/CL/kernels/CLScharr3x3Kernel.h"
-#include "arm_compute/core/CL/kernels/CLSelectKernel.h"
-#include "arm_compute/core/CL/kernels/CLSobel3x3Kernel.h"
-#include "arm_compute/core/CL/kernels/CLSobel5x5Kernel.h"
-#include "arm_compute/core/CL/kernels/CLSobel7x7Kernel.h"
-#include "arm_compute/core/CL/kernels/CLSoftmaxLayerKernel.h"
-#include "arm_compute/core/CL/kernels/CLSpaceToBatchLayerKernel.h"
-#include "arm_compute/core/CL/kernels/CLSpaceToDepthLayerKernel.h"
-#include "arm_compute/core/CL/kernels/CLStackLayerKernel.h"
-#include "arm_compute/core/CL/kernels/CLStridedSliceKernel.h"
-#include "arm_compute/core/CL/kernels/CLTableLookupKernel.h"
-#include "arm_compute/core/CL/kernels/CLThresholdKernel.h"
-#include "arm_compute/core/CL/kernels/CLTileKernel.h"
-#include "arm_compute/core/CL/kernels/CLTransposeKernel.h"
-#include "arm_compute/core/CL/kernels/CLUpsampleLayerKernel.h"
-#include "arm_compute/core/CL/kernels/CLWarpAffineKernel.h"
-#include "arm_compute/core/CL/kernels/CLWarpPerspectiveKernel.h"
-#include "arm_compute/core/CL/kernels/CLWeightsReshapeKernel.h"
-#include "arm_compute/core/CL/kernels/CLWidthConcatenate2TensorsKernel.h"
-#include "arm_compute/core/CL/kernels/CLWidthConcatenate4TensorsKernel.h"
-#include "arm_compute/core/CL/kernels/CLWidthConcatenateLayerKernel.h"
-#include "arm_compute/core/CL/kernels/CLWinogradFilterTransformKernel.h"
-#include "arm_compute/core/CL/kernels/CLWinogradInputTransformKernel.h"
-#include "arm_compute/core/CL/kernels/CLWinogradOutputTransformKernel.h"
-#include "arm_compute/core/CL/kernels/CLYOLOLayerKernel.h"
-#include "arm_compute/core/CL/kernels/ICLDepthwiseConvolutionLayer3x3Kernel.h"
-
-#endif /* ARM_COMPUTE_CLKERNELS_H */
diff --git a/arm_compute/core/CL/CLTypes.h b/arm_compute/core/CL/CLTypes.h
index c44e2c4f3f..0f6eb0dfa4 100644
--- a/arm_compute/core/CL/CLTypes.h
+++ b/arm_compute/core/CL/CLTypes.h
@@ -75,5 +75,40 @@ struct CLQuantization
     const ICLFloatArray *scale;  /**< Quantization scale array */
     const ICLInt32Array *offset; /**< Quantization offset array */
 };
+
+/** Internal keypoint structure for Lucas-Kanade Optical Flow */
+struct CLLKInternalKeypoint
+{
+    float x{ 0.f };               /**< x coordinate of the keypoint */
+    float y{ 0.f };               /**< y coordinate of the keypoint */
+    float tracking_status{ 0.f }; /**< the tracking status of the keypoint */
+    float dummy{ 0.f };           /**< Dummy field, to make sure the data structure 128-bit align, so that GPU can use vload4 */
+};
+
+/** Structure for storing Spatial Gradient Matrix and the minimum eigenvalue for each keypoint */
+struct CLCoefficientTable
+{
+    float A11;     /**< iA11 * FLT_SCALE */
+    float A12;     /**< iA11 * FLT_SCALE */
+    float A22;     /**< iA11 * FLT_SCALE */
+    float min_eig; /**< Minimum eigenvalue */
+};
+
+/** Structure for storing ival, ixval and iyval for each point inside the window */
+struct CLOldValue
+{
+    int16_t ival;  /**< ival extracts from old image */
+    int16_t ixval; /**< ixval extracts from scharr Gx image */
+    int16_t iyval; /**< iyval extracts from scharr Gy image */
+    int16_t dummy; /**< Dummy field, to make sure the data structure 128-bit align, so that GPU can use vload4 */
+};
+
+/** Interface for OpenCL Array of Internal Key Points. */
+using ICLLKInternalKeypointArray = ICLArray<CLLKInternalKeypoint>;
+/** Interface for OpenCL Array of Coefficient Tables. */
+using ICLCoefficientTableArray = ICLArray<CLCoefficientTable>;
+/** Interface for OpenCL Array of Old Values. */
+using ICLOldValArray = ICLArray<CLOldValue>;
+
 } // namespace arm_compute
 #endif /* ARM_COMPUTE_CL_TYPES_H */
diff --git a/arm_compute/core/CL/ICLKernel.h b/arm_compute/core/CL/ICLKernel.h
deleted file mode 100644
index a24cd8c798..0000000000
--- a/arm_compute/core/CL/ICLKernel.h
+++ /dev/null
@@ -1,404 +0,0 @@
-/*
- * Copyright (c) 2016-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_ICLKERNEL_H
-#define ARM_COMPUTE_ICLKERNEL_H
-
-#include "arm_compute/core/CL/CLKernelLibrary.h"
-#include "arm_compute/core/CL/CLTypes.h"
-#include "arm_compute/core/CL/OpenCL.h"
-#include "arm_compute/core/GPUTarget.h"
-#include "arm_compute/core/IKernel.h"
-#include "arm_compute/core/Validate.h"
-#include "arm_compute/core/experimental/Types.h"
-
-#include <string>
-
-namespace arm_compute
-{
-template <typename T>
-class ICLArray;
-class ICLTensor;
-class Window;
-
-/** Common interface for all the OpenCL kernels */
-class ICLKernel : public IKernel
-{
-private:
-    /** Returns the number of arguments enqueued per array object.
-     *
-     * @return The number of arguments enqueued per array object.
-     */
-    template <unsigned int        dimension_size>
-    constexpr static unsigned int num_arguments_per_array()
-    {
-        return num_arguments_per_tensor<dimension_size>();
-    }
-    /** Returns the number of arguments enqueued per tensor object.
-     *
-     * @return The number of arguments enqueued per tensor object.
-     */
-    template <unsigned int        dimension_size>
-    constexpr static unsigned int num_arguments_per_tensor()
-    {
-        return 2 + 2 * dimension_size;
-    }
-    using IKernel::configure; //Prevent children from calling IKernel::configure() directly
-protected:
-    /** Configure the kernel's window and local workgroup size hint.
-     *
-     * @param[in] window   The maximum window which will be returned by window()
-     * @param[in] lws_hint (Optional) Local-Workgroup-Size to use.
-     */
-    void configure_internal(const Window &window, cl::NDRange lws_hint = CLKernelLibrary::get().default_ndrange())
-    {
-        _lws_hint = lws_hint;
-        IKernel::configure(window);
-    }
-
-public:
-    /** Constructor */
-    ICLKernel()
-        : _kernel(nullptr), _target(GPUTarget::MIDGARD), _config_id(arm_compute::default_config_id), _max_workgroup_size(0), _lws_hint()
-    {
-    }
-    /** Returns a reference to the OpenCL kernel of this object.
-     *
-     * @return A reference to the OpenCL kernel of this object.
-     */
-    cl::Kernel &kernel()
-    {
-        return _kernel;
-    }
-    /** Add the passed 1D array's parameters to the object's kernel's arguments starting from the index idx.
-     *
-     * @param[in,out] idx            Index at which to start adding the array's arguments. Will be incremented by the number of kernel arguments set.
-     * @param[in]     array          Array to set as an argument of the object's kernel.
-     * @param[in]     strides        @ref Strides object containing stride of each dimension in bytes.
-     * @param[in]     num_dimensions Number of dimensions of the @p array.
-     * @param[in]     window         Window the kernel will be executed on.
-     */
-    template <typename T>
-    void add_1D_array_argument(unsigned int &idx, const ICLArray<T> *array, const Strides &strides, unsigned int num_dimensions, const Window &window)
-    {
-        add_array_argument<T, 1>(idx, array, strides, num_dimensions, window);
-    }
-    /** Add the passed 1D tensor's parameters to the object's kernel's arguments starting from the index idx.
-     *
-     * @param[in,out] idx    Index at which to start adding the tensor's arguments. Will be incremented by the number of kernel arguments set.
-     * @param[in]     tensor Tensor to set as an argument of the object's kernel.
-     * @param[in]     window Window the kernel will be executed on.
-     */
-    void add_1D_tensor_argument(unsigned int &idx, const ICLTensor *tensor, const Window &window)
-    {
-        add_tensor_argument<1>(idx, tensor, window);
-    }
-    /** Add the passed 1D tensor's parameters to the object's kernel's arguments starting from the index idx if the condition is true.
-     *
-     * @param[in]     cond   Condition to check
-     * @param[in,out] idx    Index at which to start adding the tensor's arguments. Will be incremented by the number of kernel arguments set.
-     * @param[in]     tensor Tensor to set as an argument of the object's kernel.
-     * @param[in]     window Window the kernel will be executed on.
-     */
-    void add_1D_tensor_argument_if(bool cond, unsigned int &idx, const ICLTensor *tensor, const Window &window)
-    {
-        if(cond)
-        {
-            add_1D_tensor_argument(idx, tensor, window);
-        }
-    }
-    /** Add the passed 2D tensor's parameters to the object's kernel's arguments starting from the index idx.
-     *
-     * @param[in,out] idx    Index at which to start adding the tensor's arguments. Will be incremented by the number of kernel arguments set.
-     * @param[in]     tensor Tensor to set as an argument of the object's kernel.
-     * @param[in]     window Window the kernel will be executed on.
-     */
-    void add_2D_tensor_argument(unsigned int &idx, const ICLTensor *tensor, const Window &window)
-    {
-        add_tensor_argument<2>(idx, tensor, window);
-    }
-    /** Add the passed 2D tensor's parameters to the object's kernel's arguments starting from the index idx if the condition is true.
-     *
-     * @param[in]     cond   Condition to check
-     * @param[in,out] idx    Index at which to start adding the tensor's arguments. Will be incremented by the number of kernel arguments set.
-     * @param[in]     tensor Tensor to set as an argument of the object's kernel.
-     * @param[in]     window Window the kernel will be executed on.
-     */
-    void add_2D_tensor_argument_if(bool cond, unsigned int &idx, const ICLTensor *tensor, const Window &window)
-    {
-        if(cond)
-        {
-            add_2D_tensor_argument(idx, tensor, window);
-        }
-    }
-    /** Add the passed 3D tensor's parameters to the object's kernel's arguments starting from the index idx.
-     *
-     * @param[in,out] idx    Index at which to start adding the tensor's arguments. Will be incremented by the number of kernel arguments set.
-     * @param[in]     tensor Tensor to set as an argument of the object's kernel.
-     * @param[in]     window Window the kernel will be executed on.
-     */
-    void add_3D_tensor_argument(unsigned int &idx, const ICLTensor *tensor, const Window &window)
-    {
-        add_tensor_argument<3>(idx, tensor, window);
-    }
-    /** Add the passed 4D tensor's parameters to the object's kernel's arguments starting from the index idx.
-     *
-     * @param[in,out] idx    Index at which to start adding the tensor's arguments. Will be incremented by the number of kernel arguments set.
-     * @param[in]     tensor Tensor to set as an argument of the object's kernel.
-     * @param[in]     window Window the kernel will be executed on.
-     */
-    void add_4D_tensor_argument(unsigned int &idx, const ICLTensor *tensor, const Window &window)
-    {
-        add_tensor_argument<4>(idx, tensor, window);
-    }
-    /** Returns the number of arguments enqueued per 1D array object.
-     *
-     * @return The number of arguments enqueues per 1D array object.
-     */
-    constexpr static unsigned int num_arguments_per_1D_array()
-    {
-        return num_arguments_per_array<1>();
-    }
-    /** Returns the number of arguments enqueued per 1D tensor object.
-     *
-     * @return The number of arguments enqueues per 1D tensor object.
-     */
-    constexpr static unsigned int num_arguments_per_1D_tensor()
-    {
-        return num_arguments_per_tensor<1>();
-    }
-    /** Returns the number of arguments enqueued per 2D tensor object.
-     *
-     * @return The number of arguments enqueues per 2D tensor object.
-     */
-    constexpr static unsigned int num_arguments_per_2D_tensor()
-    {
-        return num_arguments_per_tensor<2>();
-    }
-    /** Returns the number of arguments enqueued per 3D tensor object.
-     *
-     * @return The number of arguments enqueues per 3D tensor object.
-     */
-    constexpr static unsigned int num_arguments_per_3D_tensor()
-    {
-        return num_arguments_per_tensor<3>();
-    }
-    /** Returns the number of arguments enqueued per 4D tensor object.
-     *
-     * @return The number of arguments enqueues per 4D tensor object.
-     */
-    constexpr static unsigned int num_arguments_per_4D_tensor()
-    {
-        return num_arguments_per_tensor<4>();
-    }
-    /** Enqueue the OpenCL kernel to process the given window  on the passed OpenCL command queue.
-     *
-     * @note The queue is *not* flushed by this method, and therefore the kernel will not have been executed by the time this method returns.
-     *
-     * @param[in]     window Region on which to execute the kernel. (Must be a valid region of the window returned by window()).
-     * @param[in,out] queue  Command queue on which to enqueue the kernel.
-     */
-    virtual void run(const Window &window, cl::CommandQueue &queue)
-    {
-        ARM_COMPUTE_UNUSED(window, queue);
-    }
-    /** Enqueue the OpenCL kernel to process the given window  on the passed OpenCL command queue.
-     *
-     * @note The queue is *not* flushed by this method, and therefore the kernel will not have been executed by the time this method returns.
-     *
-     * @param[in]     tensors A vector containing the tensors to operato on.
-     * @param[in]     window  Region on which to execute the kernel. (Must be a valid region of the window returned by window()).
-     * @param[in,out] queue   Command queue on which to enqueue the kernel.
-     */
-    virtual void run_op(ITensorPack &tensors, const Window &window, cl::CommandQueue &queue)
-    {
-        ARM_COMPUTE_UNUSED(tensors, window, queue);
-    }
-    /** Add the passed parameters to the object's kernel's arguments starting from the index idx.
-     *
-     * @param[in,out] idx   Index at which to start adding the arguments. Will be incremented by the number of kernel arguments set.
-     * @param[in]     value Value to set as an argument of the object's kernel.
-     */
-    template <typename T>
-    void add_argument(unsigned int &idx, T value)
-    {
-        _kernel.setArg(idx++, value);
-    }
-
-    /** Set the Local-Workgroup-Size hint
-     *
-     * @note This method should be called after the configuration of the kernel
-     *
-     * @param[in] lws_hint Local-Workgroup-Size to use
-     */
-    void set_lws_hint(const cl::NDRange &lws_hint)
-    {
-        ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL(this); // lws_hint will be overwritten by configure()
-        _lws_hint = lws_hint;
-    }
-
-    /** Return the Local-Workgroup-Size hint
-     *
-     * @return Current lws hint
-     */
-    cl::NDRange lws_hint() const
-    {
-        return _lws_hint;
-    }
-
-    /** Get the configuration ID
-     *
-     * @note The configuration ID can be used by the caller to distinguish different calls of the same OpenCL kernel
-     *       In particular, this method can be used by CLScheduler to keep track of the best LWS for each configuration of the same kernel.
-     *       The configuration ID should be provided only for the kernels potentially affected by the LWS geometry
-     *
-     * @note This method should be called after the configuration of the kernel
-     *
-     * @return configuration id string
-     */
-    const std::string &config_id() const
-    {
-        return _config_id;
-    }
-
-    /** Set the targeted GPU architecture
-     *
-     * @param[in] target The targeted GPU architecture
-     */
-    void set_target(GPUTarget target)
-    {
-        _target = target;
-    }
-
-    /** Set the targeted GPU architecture according to the CL device
-     *
-     * @param[in] device A CL device
-     */
-    void set_target(cl::Device &device);
-
-    /** Get the targeted GPU architecture
-     *
-     * @return The targeted GPU architecture.
-     */
-    GPUTarget get_target() const
-    {
-        return _target;
-    }
-
-    /** Get the maximum workgroup size for the device the CLKernelLibrary uses.
-     *
-     * @return The maximum workgroup size value.
-     */
-    size_t get_max_workgroup_size();
-    /** Get the global work size given an execution window
-     *
-     * @param[in] window Execution window
-     *
-     * @return Global work size of the given execution window
-     */
-    static cl::NDRange gws_from_window(const Window &window);
-
-private:
-    /** Add the passed array's parameters to the object's kernel's arguments starting from the index idx.
-     *
-     * @param[in,out] idx            Index at which to start adding the array's arguments. Will be incremented by the number of kernel arguments set.
-     * @param[in]     array          Array to set as an argument of the object's kernel.
-     * @param[in]     strides        @ref Strides object containing stride of each dimension in bytes.
-     * @param[in]     num_dimensions Number of dimensions of the @p array.
-     * @param[in]     window         Window the kernel will be executed on.
-     */
-    template <typename T, unsigned int dimension_size>
-    void add_array_argument(unsigned int &idx, const ICLArray<T> *array, const Strides &strides, unsigned int num_dimensions, const Window &window);
-    /** Add the passed tensor's parameters to the object's kernel's arguments starting from the index idx.
-     *
-     * @param[in,out] idx    Index at which to start adding the tensor's arguments. Will be incremented by the number of kernel arguments set.
-     * @param[in]     tensor Tensor to set as an argument of the object's kernel.
-     * @param[in]     window Window the kernel will be executed on.
-     */
-    template <unsigned int dimension_size>
-    void add_tensor_argument(unsigned int &idx, const ICLTensor *tensor, const Window &window);
-
-protected:
-    cl::Kernel  _kernel;             /**< OpenCL kernel to run */
-    GPUTarget   _target;             /**< The targeted GPU */
-    std::string _config_id;          /**< Configuration ID */
-    size_t      _max_workgroup_size; /**< The maximum workgroup size for this kernel */
-private:
-    cl::NDRange _lws_hint; /**< Local workgroup size hint for the OpenCL kernel */
-};
-
-/** Add the kernel to the command queue with the given window.
- *
- * @note Depending on the size of the window, this might translate into several jobs being enqueued.
- *
- * @note If kernel->kernel() is empty then the function will return without adding anything to the queue.
- *
- * @param[in,out] queue                OpenCL command queue.
- * @param[in]     kernel               Kernel to enqueue
- * @param[in]     window               Window the kernel has to process.
- * @param[in]     lws_hint             (Optional) Local workgroup size requested. Default is based on the device target.
- * @param[in]     use_dummy_work_items (Optional) Use dummy work items in order to have two dimensional power of two NDRange. Default is false
- *                                     Note: it is kernel responsibility to check if the work-item is out-of-range
- *
- * @note If any dimension of the lws is greater than the global workgroup size then no lws will be passed.
- */
-void enqueue(cl::CommandQueue &queue, ICLKernel &kernel, const Window &window, const cl::NDRange &lws_hint = CLKernelLibrary::get().default_ndrange(), bool use_dummy_work_items = false);
-
-/** Add the passed array's parameters to the object's kernel's arguments starting from the index idx.
- *
- * @param[in,out] idx            Index at which to start adding the array's arguments. Will be incremented by the number of kernel arguments set.
- * @param[in]     array          Array to set as an argument of the object's kernel.
- * @param[in]     strides        @ref Strides object containing stride of each dimension in bytes.
- * @param[in]     num_dimensions Number of dimensions of the @p array.
- * @param[in]     window         Window the kernel will be executed on.
- */
-template <typename T, unsigned int dimension_size>
-void ICLKernel::add_array_argument(unsigned &idx, const ICLArray<T> *array, const Strides &strides, unsigned int num_dimensions, const Window &window)
-{
-    ARM_COMPUTE_ERROR_ON(array == nullptr);
-
-    // Calculate offset to the start of the window
-    unsigned int offset_first_element = 0;
-
-    for(unsigned int n = 0; n < num_dimensions; ++n)
-    {
-        offset_first_element += window[n].start() * strides[n];
-    }
-
-    unsigned int idx_start = idx;
-    _kernel.setArg(idx++, array->cl_buffer());
-
-    for(unsigned int dimension = 0; dimension < dimension_size; dimension++)
-    {
-        _kernel.setArg<cl_uint>(idx++, strides[dimension]);
-        _kernel.setArg<cl_uint>(idx++, strides[dimension] * window[dimension].step());
-    }
-
-    _kernel.setArg<cl_uint>(idx++, offset_first_element);
-
-    ARM_COMPUTE_ERROR_ON_MSG_VAR(idx_start + num_arguments_per_array<dimension_size>() != idx,
-                                 "add_%dD_array_argument() is supposed to add exactly %d arguments to the kernel", dimension_size, num_arguments_per_array<dimension_size>());
-    ARM_COMPUTE_UNUSED(idx_start);
-}
-}
-#endif /*ARM_COMPUTE_ICLKERNEL_H */
diff --git a/arm_compute/core/CL/ICLSimple2DKernel.h b/arm_compute/core/CL/ICLSimple2DKernel.h
deleted file mode 100644
index 86561cd562..0000000000
--- a/arm_compute/core/CL/ICLSimple2DKernel.h
+++ /dev/null
@@ -1,41 +0,0 @@
-/*
- * Copyright (c) 2017-2019 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_ICLSIMPLE2DKERNEL_H
-#define ARM_COMPUTE_ICLSIMPLE2DKERNEL_H
-
-#include "arm_compute/core/CL/ICLSimpleKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for simple OpenCL kernels having 1 tensor input and 1 tensor output. This interface can be used when the work-item processes a 2D tile */
-class ICLSimple2DKernel : public ICLSimpleKernel
-{
-public:
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-};
-}
-#endif /*ARM_COMPUTE_ICLSIMPLE2DKERNEL_H */
diff --git a/arm_compute/core/CL/ICLSimple3DKernel.h b/arm_compute/core/CL/ICLSimple3DKernel.h
deleted file mode 100644
index 3b4eaf7350..0000000000
--- a/arm_compute/core/CL/ICLSimple3DKernel.h
+++ /dev/null
@@ -1,43 +0,0 @@
-/*
- * Copyright (c) 2017-2019 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_ICLSIMPLE3DKERNEL_H
-#define ARM_COMPUTE_ICLSIMPLE3DKERNEL_H
-
-#include "arm_compute/core/CL/ICLSimple2DKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for simple OpenCL kernels having 1 tensor input and 1 tensor output.
- *  Both input tensor and output tensor must have at least 3 dimensions.
- */
-class ICLSimple3DKernel : public ICLSimple2DKernel
-{
-public:
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-};
-}
-#endif /*ARM_COMPUTE_ICLSIMPLE3DKERNEL_H */
diff --git a/arm_compute/core/CL/ICLSimpleKernel.h b/arm_compute/core/CL/ICLSimpleKernel.h
deleted file mode 100644
index 805342f830..0000000000
--- a/arm_compute/core/CL/ICLSimpleKernel.h
+++ /dev/null
@@ -1,66 +0,0 @@
-/*
- * Copyright (c) 2016-2019 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_ICLSIMPLEKERNEL_H
-#define ARM_COMPUTE_ICLSIMPLEKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-#include "arm_compute/core/CL/ICLTensor.h"
-#include "arm_compute/core/Helpers.h"
-
-namespace arm_compute
-{
-/** Interface for simple OpenCL kernels having 1 tensor input and 1 tensor output */
-class ICLSimpleKernel : public ICLKernel
-{
-public:
-    /** Constructor. */
-    ICLSimpleKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    ICLSimpleKernel(const ICLSimpleKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    ICLSimpleKernel &operator=(const ICLSimpleKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    ICLSimpleKernel(ICLSimpleKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    ICLSimpleKernel &operator=(ICLSimpleKernel &&) = default;
-    /** Default destructor */
-    ~ICLSimpleKernel() = default;
-
-    /** Configure the kernel
-     *
-     * @param[in]  input                             Source tensor.
-     * @param[out] output                            Destination tensor.
-     * @param[in]  num_elems_processed_per_iteration Number of processed elements per iteration.
-     * @param[in]  border_undefined                  (Optional) True if the border mode is undefined. False if it's replicate or constant.
-     * @param[in]  border_size                       (Optional) Size of the border.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, unsigned int num_elems_processed_per_iteration, bool border_undefined = false, const BorderSize &border_size = BorderSize());
-
-protected:
-    const ICLTensor *_input;
-    ICLTensor       *_output;
-};
-}
-
-#endif /*ARM_COMPUTE_ICLSIMPLEKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLAbsoluteDifferenceKernel.h b/arm_compute/core/CL/kernels/CLAbsoluteDifferenceKernel.h
deleted file mode 100644
index f62855cbb9..0000000000
--- a/arm_compute/core/CL/kernels/CLAbsoluteDifferenceKernel.h
+++ /dev/null
@@ -1,79 +0,0 @@
-/*
- * Copyright (c) 2016-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLABSOLUTEDIFFERENCEKERNEL_H
-#define ARM_COMPUTE_CLABSOLUTEDIFFERENCEKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the absolute difference kernel.
- *
- * Absolute difference is computed by:
- * @f[ output(x,y) = | input1(x,y) - input2(x,y) | @f]
- */
-class CLAbsoluteDifferenceKernel : public ICLKernel
-{
-public:
-    /** Default constructor. */
-    CLAbsoluteDifferenceKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLAbsoluteDifferenceKernel(const CLAbsoluteDifferenceKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLAbsoluteDifferenceKernel &operator=(const CLAbsoluteDifferenceKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLAbsoluteDifferenceKernel(CLAbsoluteDifferenceKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLAbsoluteDifferenceKernel &operator=(CLAbsoluteDifferenceKernel &&) = default;
-    /** Default destructor */
-    ~CLAbsoluteDifferenceKernel() = default;
-
-    /** Set the inputs and output images.
-     *
-     * @param[in]  input1 Source tensor. Data types supported: U8/S16.
-     * @param[in]  input2 Source tensor. Data types supported: U8/S16.
-     * @param[out] output Destination tensor. Data types supported: U8/S16.
-     */
-    void configure(const ICLTensor *input1, const ICLTensor *input2, ICLTensor *output);
-    /** Set the inputs and output images.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input1          Source tensor. Data types supported: U8/S16.
-     * @param[in]  input2          Source tensor. Data types supported: U8/S16.
-     * @param[out] output          Destination tensor. Data types supported: U8/S16.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input1, const ICLTensor *input2, ICLTensor *output);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input1; /**< Source tensor 1. */
-    const ICLTensor *_input2; /**< Source tensor 2. */
-    ICLTensor       *_output; /**< Destination tensor. */
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLABSOLUTEDIFFERENCEKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLAccumulateKernel.h b/arm_compute/core/CL/kernels/CLAccumulateKernel.h
deleted file mode 100644
index e067da084f..0000000000
--- a/arm_compute/core/CL/kernels/CLAccumulateKernel.h
+++ /dev/null
@@ -1,114 +0,0 @@
-/*
- * Copyright (c) 2016-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLACCUMULATEKERNEL_H
-#define ARM_COMPUTE_CLACCUMULATEKERNEL_H
-
-#include "arm_compute/core/CL/ICLSimple2DKernel.h"
-
-#include <cstdint>
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the accumulate kernel.
- *
- * Accumulation is computed by:
- * @f[ accum(x,y) = accum(x,y) + input(x,y) @f]
- */
-class CLAccumulateKernel : public ICLSimple2DKernel
-{
-public:
-    /** Set the input and accumulation tensors.
-     *
-     * @param[in]  input Source tensor. Data types supported: U8.
-     * @param[out] accum Destination tensor. Data types supported: S16.
-     */
-    void configure(const ICLTensor *input, ICLTensor *accum);
-    /** Set the input and accumulation tensors.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Source tensor. Data types supported: U8.
-     * @param[out] accum           Destination tensor. Data types supported: S16.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *accum);
-};
-
-/** Interface for the accumulate weighted kernel.
- *
- * Weighted accumulation is computed:
- * @f[ accum(x,y) = (1 - \alpha)*accum(x,y) + \alpha*input(x,y) @f]
- *
- * Where @f$ 0 \le \alpha \le 1 @f$
- * Conceptually, the rounding for this is defined as:
- * @f[ output(x,y)= uint8( (1 - \alpha) * float32( int32( output(x,y) ) ) + \alpha * float32( int32( input(x,y) ) ) ) @f]
-*/
-class CLAccumulateWeightedKernel : public ICLSimple2DKernel
-{
-public:
-    /** Set the input and accumulation images, and the scale value.
-     *
-     * @param[in]     input Source tensor. Data types supported: U8.
-     * @param[in]     alpha Scalar value in the range [0, 1.0]. Data types supported: F32.
-     * @param[in,out] accum Accumulated tensor. Data types supported: U8.
-     */
-    void configure(const ICLTensor *input, float alpha, ICLTensor *accum);
-    /** Set the input and accumulation images, and the scale value.
-     *
-     * @param[in]     compile_context The compile context to be used.
-     * @param[in]     input           Source tensor. Data types supported: U8.
-     * @param[in]     alpha           Scalar value in the range [0, 1.0]. Data types supported: F32.
-     * @param[in,out] accum           Accumulated tensor. Data types supported: U8.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, float alpha, ICLTensor *accum);
-};
-
-/** Interface for the accumulate squared kernel.
- *
- * The accumulation of squares is computed:
- * @f[ accum(x,y) = saturate_{int16} ( (uint16) accum(x,y) + (((uint16)(input(x,y)^2)) >> (shift)) ) @f]
- *
- * Where @f$ 0 \le shift \le 15 @f$
-*/
-class CLAccumulateSquaredKernel : public ICLSimple2DKernel
-{
-public:
-    /** Set the input and accumulation tensors and the shift value.
-     *
-     * @param[in]     input Source tensor. Data types supported: U8.
-     * @param[in]     shift Shift value in the range of [0, 15]. Data types supported: U32.
-     * @param[in,out] accum Accumulated tensor. Data types supported: S16.
-     */
-    void configure(const ICLTensor *input, uint32_t shift, ICLTensor *accum);
-    /** Set the input and accumulation tensors and the shift value.
-     *
-     * @param[in]     compile_context The compile context to be used.
-     * @param[in]     input           Source tensor. Data types supported: U8.
-     * @param[in]     shift           Shift value in the range of [0, 15]. Data types supported: U32.
-     * @param[in,out] accum           Accumulated tensor. Data types supported: S16.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, uint32_t shift, ICLTensor *accum);
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLACCUMULATEKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLActivationLayerKernel.h b/arm_compute/core/CL/kernels/CLActivationLayerKernel.h
deleted file mode 100644
index 81d4ccb065..0000000000
--- a/arm_compute/core/CL/kernels/CLActivationLayerKernel.h
+++ /dev/null
@@ -1,77 +0,0 @@
-/*
- * Copyright (c) 2016-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLACTIVATIONLAYERKERNEL_H
-#define ARM_COMPUTE_CLACTIVATIONLAYERKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-/** Interface for the activation layer kernel. */
-class CLActivationLayerKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLActivationLayerKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLActivationLayerKernel(const CLActivationLayerKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLActivationLayerKernel &operator=(const CLActivationLayerKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLActivationLayerKernel(CLActivationLayerKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLActivationLayerKernel &operator=(CLActivationLayerKernel &&) = default;
-    /** Default destructor */
-    ~CLActivationLayerKernel() = default;
-    /** Set the input and output tensor.
-     *
-     * @note If the output tensor is a nullptr, the activation function will be performed in-place
-     *
-     * @param[in]      compile_context The compile context to be used.
-     * @param[in, out] input           Source tensor. In case of @p output tensor = nullptr, this tensor will store the result
-     *                                 of the activation function. Data types supported: QASYMM8/QASYMM8_SIGNED/QSYMM16/F16/F32.
-     * @param[out]     output          Destination tensor. Data type supported: same as @p input
-     * @param[in]      act_info        Activation layer information.
-     */
-    void configure(const CLCompileContext &compile_context, ITensorInfo *input, ITensorInfo *output, ActivationLayerInfo act_info);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLActivationLayerKernel
-     *
-     * @param[in] input    Source tensor info. In case of @p output tensor info = nullptr, this tensor will store the result
-     *                     of the activation function. Data types supported: QASYMM8/QASYMM8_SIGNED/QSYMM16/F16/F32.
-     * @param[in] output   Destination tensor info. Data type supported: same as @p input
-     * @param[in] act_info Activation layer information.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output, const ActivationLayerInfo &act_info);
-
-    // Inherited methods overridden:
-    void run_op(ITensorPack &tensors, const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    bool _run_in_place;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLACTIVATIONLAYERKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLArgMinMaxLayerKernel.h b/arm_compute/core/CL/kernels/CLArgMinMaxLayerKernel.h
deleted file mode 100644
index 48876c0b56..0000000000
--- a/arm_compute/core/CL/kernels/CLArgMinMaxLayerKernel.h
+++ /dev/null
@@ -1,106 +0,0 @@
-/*
- * Copyright (c) 2019-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLARGMINMAXLAYERKERNEL_H
-#define ARM_COMPUTE_CLARGMINMAXLAYERKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-#include "arm_compute/core/Types.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the reduction operation kernel
- *
- * @note The default data type for an uninitialized output tensor is
- *       signed 32-bit integer (S32). It is the user's responsibility to check
- *       that the results do not overflow because the indices are computed
- *       in unsigned 32-bit (U32).
- */
-class CLArgMinMaxLayerKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLArgMinMaxLayerKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLArgMinMaxLayerKernel(const CLArgMinMaxLayerKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLArgMinMaxLayerKernel &operator=(const CLArgMinMaxLayerKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLArgMinMaxLayerKernel(CLArgMinMaxLayerKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLArgMinMaxLayerKernel &operator=(CLArgMinMaxLayerKernel &&) = default;
-    /** Default destructor */
-    ~CLArgMinMaxLayerKernel() = default;
-
-    /** Set the input and output tensors.
-     *
-     * @param[in]  input       Source tensor. Data types supported: QASYMM8/QASYMM8_SIGNED/S32/F16/F32.
-     * @param[in]  prev_output Destination tensor of the previous iterations of @ref CLArgMinMaxLayerKernel. Data types supported: U32/S32
-     *                         Has to be nullptr for the first iteration
-     * @param[out] output      Destination tensor. Data types supported: U32/S32
-     *                         Output will have the same number of dimensions as input.
-     * @param[in]  axis        Axis along which to reduce. Supported reduction axis : 0,1,2,3
-     * @param[in]  op          Reduction operation to perform. Only ArgMin and ArgMax are supported.
-     */
-    void configure(const ICLTensor *input, const ICLTensor *prev_output, ICLTensor *output, unsigned int axis, ReductionOperation op);
-    /** Set the input and output tensors.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Source tensor. Data types supported: QASYMM8/QASYMM8_SIGNED/S32/F16/F32.
-     * @param[in]  prev_output     Destination tensor of the previous iterations of @ref CLArgMinMaxLayerKernel. Data types supported: U32/S32
-     *                             Has to be nullptr for the first iteration
-     * @param[out] output          Destination tensor. Data types supported: U32/S32
-     *                             Output will have the same number of dimensions as input.
-     * @param[in]  axis            Axis along which to reduce. Supported reduction axis : 0,1,2,3
-     * @param[in]  op              Reduction operation to perform. Only ArgMin and ArgMax are supported.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, const ICLTensor *prev_output, ICLTensor *output, unsigned int axis, ReductionOperation op);
-
-    /** Static function to check if given info will lead to a valid configuration of @ref CLArgMinMaxLayerKernel.
-     *
-     * @param[in] input       Source tensor info. Data types supported: QASYMM8/QASYMM8_SIGNED/S32/F16/F32.
-     * @param[in] prev_output Destination tensor info of the previous iterations. Data types supported: U32/S32
-     *                        Has to be nullptr for the first iteration
-     * @param[in] output      Destination tensor info. Data types supported: U32/S32
-     *                        Output will have the same number of dimensions as input.
-     * @param[in] axis        Axis along which to reduce. Supported reduction axis : 0,1,2,3
-     * @param[in] op          Reduction operation to perform.  Only ArgMin and ArgMax are supported.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *prev_output, const ITensorInfo *output, unsigned int axis, ReductionOperation op);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor   *_input;
-    const ICLTensor   *_prev_output;
-    ICLTensor         *_output;
-    unsigned int       _reduction_axis;
-    ReductionOperation _op;
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLARGMINMAXLAYERKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLBatchConcatenateLayerKernel.h b/arm_compute/core/CL/kernels/CLBatchConcatenateLayerKernel.h
deleted file mode 100644
index bb8968ca83..0000000000
--- a/arm_compute/core/CL/kernels/CLBatchConcatenateLayerKernel.h
+++ /dev/null
@@ -1,82 +0,0 @@
-/*
- * Copyright (c) 2019-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-
-#ifndef ARM_COMPUTE_CLBATCHCONCATENATEKERNEL_H
-#define ARM_COMPUTE_CLBATCHCONCATENATEKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-#include "arm_compute/core/Types.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the batch concatenate kernel.
- *  The input tensor will be concatenated into the output tensor.
- */
-class CLBatchConcatenateLayerKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLBatchConcatenateLayerKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLBatchConcatenateLayerKernel(const CLBatchConcatenateLayerKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLBatchConcatenateLayerKernel &operator=(const CLBatchConcatenateLayerKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLBatchConcatenateLayerKernel(CLBatchConcatenateLayerKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLBatchConcatenateLayerKernel &operator=(CLBatchConcatenateLayerKernel &&) = default;
-    /** Default destructor */
-    ~CLBatchConcatenateLayerKernel() = default;
-    /** Initialise the kernel's inputs and output
-     *
-     * @param[in]     compile_context The compile context to be used.
-     * @param[in]     input           Input tensor. Data types supported: All.
-     * @param[in]     batch_offset    The offset on axis # 3.
-     * @param[in,out] output          Output tensor. Data types supported: Same as @p input.
-     *
-     * @note: The output tensor's low two dimensions can't be smaller than the input one's.
-     * @note: The gaps between the two lowest dimensions of input and output need to be divisible by 2.
-     *
-     */
-    void configure(const CLCompileContext &compile_context, ITensorInfo *input, unsigned int batch_offset, ITensorInfo *output);
-    /**  Static function to check if given info will lead to a valid configuration of @ref CLBatchConcatenateLayerKernel
-     *
-     * @param[in] input        Input tensor info. Data types supported: All.
-     * @param[in] batch_offset The offset on axis # 3.
-     * @param[in] output       Output tensor info. Data types supported: Same as @p input.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, unsigned int batch_offset, const ITensorInfo *output);
-
-    // Inherited methods overridden:
-    void run_op(ITensorPack &tensors, const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    unsigned int _batch_offset;
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLBATCHCONCATENATEKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLBatchNormalizationLayerKernel.h b/arm_compute/core/CL/kernels/CLBatchNormalizationLayerKernel.h
deleted file mode 100644
index c556a0c6f4..0000000000
--- a/arm_compute/core/CL/kernels/CLBatchNormalizationLayerKernel.h
+++ /dev/null
@@ -1,120 +0,0 @@
-/*
- * Copyright (c) 2017-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLBATCHNORMALIZATIONLAYERKERNEL_H
-#define ARM_COMPUTE_CLBATCHNORMALIZATIONLAYERKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the BatchNormalization layer kernel.
- */
-class CLBatchNormalizationLayerKernel : public ICLKernel
-{
-public:
-    /** Constructor */
-    CLBatchNormalizationLayerKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLBatchNormalizationLayerKernel(const CLBatchNormalizationLayerKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLBatchNormalizationLayerKernel &operator=(const CLBatchNormalizationLayerKernel &) = delete;
-    /** Default Move Constructor. */
-    CLBatchNormalizationLayerKernel(CLBatchNormalizationLayerKernel &&) = default;
-    /** Default move assignment operator */
-    CLBatchNormalizationLayerKernel &operator=(CLBatchNormalizationLayerKernel &&) = default;
-    /** Default destructor */
-    ~CLBatchNormalizationLayerKernel() = default;
-
-    /** Set the input and output tensors.
-     *
-     * @note If the output tensor is a nullptr, the batch normalization function will be performed in-place
-     *
-     * @param[in, out] input    Source tensor. In case of @p output tensor = nullptr, this tensor will store the result.
-     *                          3 lower dimensions represent a single input with dimensions [width, height, FM].
-     *                          The rest are optional and used for representing batches. Data types supported: F16/F32. Data layouts supported: NCHW/NHWC.
-     * @param[out]     output   Destination tensor. Output will have the same number of dimensions as input. Data type supported: same as @p input
-     * @param[in]      mean     Mean values tensor. 1 dimension with size equal to the feature maps [FM]. Data types supported: Same as @p input
-     * @param[in]      var      Variance values tensor. 1 dimension with size equal to the feature maps [FM]. Data types supported: Same as @p input
-     * @param[in]      beta     (Optional) Beta values tensor info. 1 dimension with size equal to the feature maps [FM]. If not provided, default value for beta is 0. Data types supported: Same as @p input
-     * @param[in]      gamma    (Optional) Gamma values tensor info. 1 dimension with size equal to the feature maps [FM]. If not provided, default value for gamma is 1. Data types supported: Same as @p input
-     * @param[in]      epsilon  (Optional) Small value to avoid division with zero. Default value is 0.001f.
-     * @param[in]      act_info (Optional) Activation layer information in case of a fused activation. Only RELU, BOUNDED_RELU and LU_BOUNDED_RELU supported.
-     */
-    void configure(ICLTensor *input, ICLTensor *output, const ICLTensor *mean, const ICLTensor *var, const ICLTensor *beta = nullptr, const ICLTensor *gamma = nullptr, float epsilon = 0.001f,
-                   ActivationLayerInfo act_info = ActivationLayerInfo());
-    /** Set the input and output tensors.
-     *
-     * @note If the output tensor is a nullptr, the batch normalization function will be performed in-place
-     *
-     * @param[in]      compile_context The compile context to be used.
-     * @param[in, out] input           Source tensor. In case of @p output tensor = nullptr, this tensor will store the result.
-     *                                 3 lower dimensions represent a single input with dimensions [width, height, FM].
-     *                                 The rest are optional and used for representing batches. Data types supported: F16/F32. Data layouts supported: NCHW/NHWC.
-     * @param[out]     output          Destination tensor. Output will have the same number of dimensions as input. Data type supported: same as @p input
-     * @param[in]      mean            Mean values tensor. 1 dimension with size equal to the feature maps [FM]. Data types supported: Same as @p input
-     * @param[in]      var             Variance values tensor. 1 dimension with size equal to the feature maps [FM]. Data types supported: Same as @p input
-     * @param[in]      beta            (Optional) Beta values tensor info. 1 dimension with size equal to the feature maps [FM]. If not provided, default value for beta is 0. Data types supported: Same as @p input
-     * @param[in]      gamma           (Optional) Gamma values tensor info. 1 dimension with size equal to the feature maps [FM]. If not provided, default value for gamma is 1. Data types supported: Same as @p input
-     * @param[in]      epsilon         (Optional) Small value to avoid division with zero. Default value is 0.001f.
-     * @param[in]      act_info        (Optional) Activation layer information in case of a fused activation. Only RELU, BOUNDED_RELU and LU_BOUNDED_RELU supported.
-     */
-    void configure(const CLCompileContext &compile_context, ICLTensor *input, ICLTensor *output, const ICLTensor *mean, const ICLTensor *var, const ICLTensor *beta = nullptr,
-                   const ICLTensor *gamma = nullptr, float epsilon = 0.001f, ActivationLayerInfo act_info = ActivationLayerInfo());
-    /** Static function to check if given info will lead to a valid configuration of @ref CLBatchNormalizationLayerKernel
-     *
-     * @param[in] input    Source tensor info. In case of @p output tensor info = nullptr, this tensor will store the result.
-     *                     3 lower dimensions represent a single input with dimensions [width, height, FM].
-     *                     The rest are optional and used for representing batches. Data types supported: F16/F32. Data layouts supported: NCHW/NHWC.
-     * @param[in] output   Destination tensor info. Output will have the same number of dimensions as input. Data type supported: same as @p input
-     * @param[in] mean     Mean values tensor info. 1 dimension with size equal to the feature maps [FM]. Data types supported: Same as @p input
-     * @param[in] var      Variance values tensor info. 1 dimension with size equal to the feature maps [FM]. Data types supported: Same as @p input
-     * @param[in] beta     (Optional) Beta values tensor info. 1 dimension with size equal to the feature maps [FM]. If not provided, default value for beta is 0. Data types supported: Same as @p input
-     * @param[in] gamma    (Optional) Gamma values tensor info. 1 dimension with size equal to the feature maps [FM]. If not provided, default value for gamma is 1. Data types supported: Same as @p input
-     * @param[in] epsilon  (Optional) Small value to avoid division with zero. Default value is 0.001f.
-     * @param[in] act_info (Optional) Activation layer information in case of a fused activation. Only RELU, BOUNDED_RELU and LU_BOUNDED_RELU supported.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output,
-                           const ITensorInfo *mean, const ITensorInfo *var,
-                           const ITensorInfo *beta = nullptr, const ITensorInfo *gamma = nullptr,
-                           float epsilon = 0.001f, ActivationLayerInfo act_info = ActivationLayerInfo());
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    ICLTensor       *_input;
-    ICLTensor       *_output;
-    const ICLTensor *_mean;
-    const ICLTensor *_var;
-    const ICLTensor *_beta;
-    const ICLTensor *_gamma;
-    float            _epsilon;
-    bool             _run_in_place;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLBATCHNORMALIZATIONLAYERKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLBatchToSpaceLayerKernel.h b/arm_compute/core/CL/kernels/CLBatchToSpaceLayerKernel.h
deleted file mode 100644
index 7af88d8986..0000000000
--- a/arm_compute/core/CL/kernels/CLBatchToSpaceLayerKernel.h
+++ /dev/null
@@ -1,111 +0,0 @@
-/*
- * Copyright (c) 2018-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLBATCHTOSPACELAYERKERNEL_H
-#define ARM_COMPUTE_CLBATCHTOSPACELAYERKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-#include "arm_compute/core/Types.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the batch to space kernel */
-class CLBatchToSpaceLayerKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLBatchToSpaceLayerKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLBatchToSpaceLayerKernel(const CLBatchToSpaceLayerKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLBatchToSpaceLayerKernel &operator=(const CLBatchToSpaceLayerKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLBatchToSpaceLayerKernel(CLBatchToSpaceLayerKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLBatchToSpaceLayerKernel &operator=(CLBatchToSpaceLayerKernel &&) = default;
-    /** Default destructor */
-    ~CLBatchToSpaceLayerKernel() = default;
-    /** Initialise the kernel's inputs and output.
-     *
-     * @param[in]  input       Tensor input. Supported tensor rank: 4. Data types supported: All.
-     * @param[in]  block_shape 1-D tensor with shape [M]. Data types supported: S32
-     * @param[out] output      Tensor output. Data types supported: same as @p input
-     */
-    void configure(const ICLTensor *input, const ICLTensor *block_shape, ICLTensor *output);
-    /** Initialise the kernel's inputs and output.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Tensor input. Supported tensor rank: 4. Data types supported: All.
-     * @param[in]  block_shape     1-D tensor with shape [M]. Data types supported: S32
-     * @param[out] output          Tensor output. Data types supported: same as @p input
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, const ICLTensor *block_shape, ICLTensor *output);
-    /** Initialise the kernel's inputs and output (Static block shape).
-     *
-     * @param[in]  input         Tensor input. Supported tensor rank: 4. Data types supported: All.
-     * @param[in]  block_shape_x Block shape x value.
-     * @param[in]  block_shape_y Block shape y value.
-     * @param[out] output        Tensor output. Data types supported: same as @p input
-     */
-    void configure(const ICLTensor *input, const int32_t block_shape_x, const int32_t block_shape_y, ICLTensor *output);
-    /** Initialise the kernel's inputs and output (Static block shape).
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Tensor input. Supported tensor rank: 4. Data types supported: All.
-     * @param[in]  block_shape_x   Block shape x value.
-     * @param[in]  block_shape_y   Block shape y value.
-     * @param[out] output          Tensor output. Data types supported: same as @p input
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, const int32_t block_shape_x, const int32_t block_shape_y, ICLTensor *output);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLBatchToSpaceLayerKernel
-     *
-     * @param[in] input       Tensor input. Supported tensor rank: 4. Data types supported: All.
-     * @param[in] block_shape 1-D tensor with shape [M]. Data types supported: S32
-     * @param[in] output      Tensor output. Data types supported: same as @p input
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *block_shape, const ITensorInfo *output);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLBatchToSpaceLayerKernel (Static block shape).
-     *
-     * @param[in] input         Tensor input. Supported tensor rank: 4. Data types supported: All.
-     * @param[in] block_shape_x Block shape x value.
-     * @param[in] block_shape_y Block shape y value.
-     * @param[in] output        Tensor output. Data types supported: same as @p input
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const int32_t block_shape_x, const int32_t block_shape_y, const ITensorInfo *output);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;       /**< Source tensor */
-    const ICLTensor *_block_shape; /**< Block shape tensor */
-    ICLTensor       *_output;      /**< Destination tensor */
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLBATCHTOSPACELAYERKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLBitwiseAndKernel.h b/arm_compute/core/CL/kernels/CLBitwiseAndKernel.h
deleted file mode 100644
index e291f08b9a..0000000000
--- a/arm_compute/core/CL/kernels/CLBitwiseAndKernel.h
+++ /dev/null
@@ -1,76 +0,0 @@
-/*
- * Copyright (c) 2016-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLBITWISEANDKERNEL_H
-#define ARM_COMPUTE_CLBITWISEANDKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the bitwise AND operation kernel.
- *
- * Result is computed by:
- * @f[ output(x,y) = input1(x,y) \land input2(x,y) @f]
- */
-class CLBitwiseAndKernel : public ICLKernel
-{
-public:
-    /** Default constructor. */
-    CLBitwiseAndKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLBitwiseAndKernel(const CLBitwiseAndKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLBitwiseAndKernel &operator=(const CLBitwiseAndKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLBitwiseAndKernel(CLBitwiseAndKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLBitwiseAndKernel &operator=(CLBitwiseAndKernel &&) = default;
-    /** Set the inputs and output images
-     *
-     * @param[in]  input1 Source tensor. Data types supported: U8.
-     * @param[in]  input2 Source tensor. Data types supported: U8.
-     * @param[out] output Destination tensor. Data types supported: U8.
-     */
-    void configure(const ICLTensor *input1, const ICLTensor *input2, ICLTensor *output);
-    /** Set the inputs and output images
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input1          Source tensor. Data types supported: U8.
-     * @param[in]  input2          Source tensor. Data types supported: U8.
-     * @param[out] output          Destination tensor. Data types supported: U8.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input1, const ICLTensor *input2, ICLTensor *output);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input1; /**< Source tensor 1 */
-    const ICLTensor *_input2; /**< Source tensor 2 */
-    ICLTensor       *_output; /**< Destination tensor */
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLBITWISEANDKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLBitwiseNotKernel.h b/arm_compute/core/CL/kernels/CLBitwiseNotKernel.h
deleted file mode 100644
index f57bbf4778..0000000000
--- a/arm_compute/core/CL/kernels/CLBitwiseNotKernel.h
+++ /dev/null
@@ -1,56 +0,0 @@
-/*
- * Copyright (c) 2016-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLBITWISENOTKERNEL_H
-#define ARM_COMPUTE_CLBITWISENOTKERNEL_H
-
-#include "arm_compute/core/CL/ICLSimple2DKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the bitwise NOT operation kernel.
- *
- * Result is computed by:
- * @f[ output(x,y) = \lnot input(x,y) @f]
- */
-class CLBitwiseNotKernel : public ICLSimple2DKernel
-{
-public:
-    /** Set the inputs and output images.
-     *
-     * @param[in]  input  Source tensor. Data types supported: U8.
-     * @param[out] output Destination tensor. Data types supported: U8.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output);
-    /** Set the inputs and output images.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Source tensor. Data types supported: U8.
-     * @param[out] output          Destination tensor. Data types supported: U8.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output);
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLBITWISENOTKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLBitwiseOrKernel.h b/arm_compute/core/CL/kernels/CLBitwiseOrKernel.h
deleted file mode 100644
index 944224ecb9..0000000000
--- a/arm_compute/core/CL/kernels/CLBitwiseOrKernel.h
+++ /dev/null
@@ -1,76 +0,0 @@
-/*
- * Copyright (c) 2016-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLBITWISEORKERNEL_H
-#define ARM_COMPUTE_CLBITWISEORKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the bitwise OR operation kernel.
- *
- * Result is computed by:
- * @f[ output(x,y) = input1(x,y) \lor input2(x,y) @f]
- */
-class CLBitwiseOrKernel : public ICLKernel
-{
-public:
-    /** Default constructor. */
-    CLBitwiseOrKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLBitwiseOrKernel(const CLBitwiseOrKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLBitwiseOrKernel &operator=(const CLBitwiseOrKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLBitwiseOrKernel(CLBitwiseOrKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLBitwiseOrKernel &operator=(CLBitwiseOrKernel &&) = default;
-    /** Set the inputs and output images
-     *
-     * @param[in]  input1 Source tensor. Data types supported: U8.
-     * @param[in]  input2 Source tensor. Data types supported: U8.
-     * @param[out] output Destination tensor. Data types supported: U8.
-     */
-    void configure(const ICLTensor *input1, const ICLTensor *input2, ICLTensor *output);
-    /** Set the inputs and output images
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input1          Source tensor. Data types supported: U8.
-     * @param[in]  input2          Source tensor. Data types supported: U8.
-     * @param[out] output          Destination tensor. Data types supported: U8.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input1, const ICLTensor *input2, ICLTensor *output);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input1; /**< Source tensor 1 */
-    const ICLTensor *_input2; /**< Source tensor 2 */
-    ICLTensor       *_output; /**< Destination tensor */
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLBITWISEORKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLBitwiseXorKernel.h b/arm_compute/core/CL/kernels/CLBitwiseXorKernel.h
deleted file mode 100644
index 732ae8659e..0000000000
--- a/arm_compute/core/CL/kernels/CLBitwiseXorKernel.h
+++ /dev/null
@@ -1,76 +0,0 @@
-/*
- * Copyright (c) 2016-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLBITWISEXORKERNEL_H
-#define ARM_COMPUTE_CLBITWISEXORKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the bitwise XOR operation kernel.
- *
- * Result is computed by:
- * @f[ output(x,y) = input1(x,y) \oplus input2(x,y) @f]
- */
-class CLBitwiseXorKernel : public ICLKernel
-{
-public:
-    /** Default constructor. */
-    CLBitwiseXorKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLBitwiseXorKernel(const CLBitwiseXorKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLBitwiseXorKernel &operator=(const CLBitwiseXorKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLBitwiseXorKernel(CLBitwiseXorKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLBitwiseXorKernel &operator=(CLBitwiseXorKernel &&) = default;
-    /** Set the inputs and output images
-     *
-     * @param[in]  input1 Source tensor. Data types supported: U8.
-     * @param[in]  input2 Source tensor. Data types supported: U8.
-     * @param[out] output Destination tensor. Data types supported: U8.
-     */
-    void configure(const ICLTensor *input1, const ICLTensor *input2, ICLTensor *output);
-    /** Set the inputs and output images
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input1          Source tensor. Data types supported: U8.
-     * @param[in]  input2          Source tensor. Data types supported: U8.
-     * @param[out] output          Destination tensor. Data types supported: U8.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input1, const ICLTensor *input2, ICLTensor *output);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input1; /**< Source tensor 1 */
-    const ICLTensor *_input2; /**< Source tensor 2 */
-    ICLTensor       *_output; /**< Destination tensor */
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLBITWISEXORKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLBoundingBoxTransformKernel.h b/arm_compute/core/CL/kernels/CLBoundingBoxTransformKernel.h
deleted file mode 100644
index 4e8c5a6f18..0000000000
--- a/arm_compute/core/CL/kernels/CLBoundingBoxTransformKernel.h
+++ /dev/null
@@ -1,99 +0,0 @@
-/*
- * Copyright (c) 2018-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLBOUNDINGBOXTRANSFORMKERNEL_H
-#define ARM_COMPUTE_CLBOUNDINGBOXTRANSFORMKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the bounding box kernel */
-class CLBoundingBoxTransformKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLBoundingBoxTransformKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLBoundingBoxTransformKernel(const CLBoundingBoxTransformKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLBoundingBoxTransformKernel &operator=(const CLBoundingBoxTransformKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLBoundingBoxTransformKernel(CLBoundingBoxTransformKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLBoundingBoxTransformKernel &operator=(CLBoundingBoxTransformKernel &&) = default;
-    /** Default destructor */
-    ~CLBoundingBoxTransformKernel() = default;
-
-    /** Set the input and output tensors.
-     *
-     * @param[in]  boxes      Source tensor. Bounding box proposals in pixel coordinates. Size(M, 4), format [x1, y1, x2, y2]. Data types supported: QASYMM16/F16/F32.
-     * @param[out] pred_boxes Destination tensor. Pixel coordinates of the transformed bounding boxes. Size (M, 4*K), format [x1, y1, x2, y2]. Data types supported: Same as @p input
-     * @param[in]  deltas     Bounding box translations and scales. Size (M, 4*K), format [dx, dy, dw, dh], K  is the number of classes.
-     *                        Data types supported: QASYMM8 if @p input is QASYMM16, otherwise same as @p input
-     * @param[in]  info       Contains BoundingBox operation information described in @ref BoundingBoxTransformInfo.
-     *
-     * @note Only single image prediction is supported. Height and Width (and scale) of the image will be contained in the BoundingBoxTransformInfo struct.
-     *
-     */
-    void configure(const ICLTensor *boxes, ICLTensor *pred_boxes, const ICLTensor *deltas, const BoundingBoxTransformInfo &info);
-    /** Set the input and output tensors.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  boxes           Source tensor. Bounding box proposals in pixel coordinates. Size(M, 4), format [x1, y1, x2, y2]. Data types supported: QASYMM16/F16/F32.
-     * @param[out] pred_boxes      Destination tensor. Pixel coordinates of the transformed bounding boxes. Size (M, 4*K), format [x1, y1, x2, y2]. Data types supported: Same as @p input
-     * @param[in]  deltas          Bounding box translations and scales. Size (M, 4*K), format [dx, dy, dw, dh], K  is the number of classes.
-     *                             Data types supported: QASYMM8 if @p input is QASYMM16, otherwise same as @p input
-     * @param[in]  info            Contains BoundingBox operation information described in @ref BoundingBoxTransformInfo.
-     *
-     * @note Only single image prediction is supported. Height and Width (and scale) of the image will be contained in the BoundingBoxTransformInfo struct.
-     *
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *boxes, ICLTensor *pred_boxes, const ICLTensor *deltas, const BoundingBoxTransformInfo &info);
-
-    /** Static function to check if given info will lead to a valid configuration of @ref CLBoundingBoxTransform
-     *
-     * @param[in] boxes      Source tensor info. Bounding box proposals in pixel coordinates. Size(M, 4), format [x1, y1, x2, y2]. Data types supported: QASYMM16/F16/F32.
-     * @param[in] pred_boxes Destination tensor info. Pixel coordinates of the transformed bounding boxes. Size (M, 4*K), format [x1, y1, x2, y2]. Data types supported: Same as @p input
-     * @param[in] deltas     Bounding box translations and scales. Size (M, 4*K), format [dx, dy, dw, dh], K  is the number of classes.
-     *                       Data types supported: QASYMM8 if @p input is QASYMM16, otherwise same as @p input
-     * @param[in] info       Contains BoundingBox operation information described in @ref BoundingBoxTransformInfo.
-     *
-     * @note Only single image prediction is supported. Height and Width (and scale) of the image will be contained in the BoundingBoxTransformInfo struct.
-     *
-     * @return a Status
-     */
-    static Status validate(const ITensorInfo *boxes, const ITensorInfo *pred_boxes, const ITensorInfo *deltas, const BoundingBoxTransformInfo &info);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_boxes;
-    ICLTensor       *_pred_boxes;
-    const ICLTensor *_deltas;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLBOUNDINGBOXTRANSFORMKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLBox3x3Kernel.h b/arm_compute/core/CL/kernels/CLBox3x3Kernel.h
deleted file mode 100644
index 1a8572dd68..0000000000
--- a/arm_compute/core/CL/kernels/CLBox3x3Kernel.h
+++ /dev/null
@@ -1,59 +0,0 @@
-/*
- * Copyright (c) 2016-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLBOX3X3KERNEL_H
-#define ARM_COMPUTE_CLBOX3X3KERNEL_H
-
-#include "arm_compute/core/CL/ICLSimple2DKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the box 3x3 filter kernel.
- *
- */
-class CLBox3x3Kernel : public ICLSimple2DKernel
-{
-public:
-    /**Initialise the kernel's input and output.
-     *
-     * @param[in]  input            An input tensor. Data types supported: U8
-     * @param[out] output           The output tensor. Data types supported: U8.
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, bool border_undefined);
-    /**Initialise the kernel's input and output.
-     *
-     * @param[in]  compile_context  The compile context to be used.
-     * @param[in]  input            An input tensor. Data types supported: U8
-     * @param[out] output           The output tensor. Data types supported: U8.
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, bool border_undefined);
-
-    //Inherited methods overriden:
-    BorderSize border_size() const override;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLBOX3X3KERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLCannyEdgeKernel.h b/arm_compute/core/CL/kernels/CLCannyEdgeKernel.h
deleted file mode 100644
index c4d0297aef..0000000000
--- a/arm_compute/core/CL/kernels/CLCannyEdgeKernel.h
+++ /dev/null
@@ -1,185 +0,0 @@
-/*
- * Copyright (c) 2017-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLCANNYEDGEKERNEL_H
-#define ARM_COMPUTE_CLCANNYEDGEKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** OpenCL kernel to perform Gradient computation.
- */
-class CLGradientKernel : public ICLKernel
-{
-public:
-    /** Constructor */
-    CLGradientKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLGradientKernel(const CLGradientKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLGradientKernel &operator=(const CLGradientKernel &) = delete;
-    /** Initialise the kernel's sources, destinations and border mode.
-     *
-     * @note gx, gy and mag must all be the same size (either 16 or 32).
-     *
-     * @param[in]  gx        Source tensor - Gx component. Data types supported: S16/S32.
-     * @param[in]  gy        Source tensor - Gy component. Data types supported: Same as gx.
-     * @param[out] magnitude Destination tensor - Magnitude. Data types supported: U16/U32. Must match the pixel size of gx, gy.
-     * @param[out] phase     Destination tensor - Quantized phase. Data types supported: U8.
-     * @param[in]  norm_type Normalization type. if 1, L1-Norm otherwise L2-Norm.
-     */
-    void configure(const ICLTensor *gx, const ICLTensor *gy, ICLTensor *magnitude, ICLTensor *phase, int32_t norm_type);
-    /** Initialise the kernel's sources, destinations and border mode.
-     *
-     * @note gx, gy and mag must all be the same size (either 16 or 32).
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  gx              Source tensor - Gx component. Data types supported: S16/S32.
-     * @param[in]  gy              Source tensor - Gy component. Data types supported: Same as gx.
-     * @param[out] magnitude       Destination tensor - Magnitude. Data types supported: U16/U32. Must match the pixel size of gx, gy.
-     * @param[out] phase           Destination tensor - Quantized phase. Data types supported: U8.
-     * @param[in]  norm_type       Normalization type. if 1, L1-Norm otherwise L2-Norm.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *gx, const ICLTensor *gy, ICLTensor *magnitude, ICLTensor *phase, int32_t norm_type);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_gx;        /**< Source tensor - Gx component */
-    const ICLTensor *_gy;        /**< Source tensor - Gy component */
-    ICLTensor       *_magnitude; /**< Destination tensor - Magnitude */
-    ICLTensor       *_phase;     /**< Destination tensor - Quantized phase */
-};
-
-/** OpenCL kernel to perform Non-Maxima suppression for Canny Edge.
- *
- * @note This kernel is meant to be used alongside CannyEdge and performs a non-maxima suppression using magnitude and phase of input
- *       to characterize points as possible edges. The output buffer needs to be cleared before this kernel is executed.
- *
- * @note Hysteresis is computed in @ref CLEdgeTraceKernel
- */
-class CLEdgeNonMaxSuppressionKernel : public ICLKernel
-{
-public:
-    /** Constructor */
-    CLEdgeNonMaxSuppressionKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLEdgeNonMaxSuppressionKernel(const CLEdgeNonMaxSuppressionKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLEdgeNonMaxSuppressionKernel &operator=(const CLEdgeNonMaxSuppressionKernel &) = delete;
-    /** Initialise the kernel's sources, destination and border mode.
-     *
-     * @param[in]  magnitude        Source tensor - Magnitude. Data types supported: U16/U32.
-     * @param[in]  phase            Source tensor - Quantized phase. Data types supported: U8.
-     * @param[out] output           Destination tensor. Data types supported: U16/U32.
-     * @param[in]  lower_thr        Lower threshold.
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const ICLTensor *magnitude, const ICLTensor *phase, ICLTensor *output, int32_t lower_thr, bool border_undefined);
-    /** Initialise the kernel's sources, destination and border mode.
-     *
-     * @param[in]  compile_context  The compile context to be used.
-     * @param[in]  magnitude        Source tensor - Magnitude. Data types supported: U16/U32.
-     * @param[in]  phase            Source tensor - Quantized phase. Data types supported: U8.
-     * @param[out] output           Destination tensor. Data types supported: U16/U32.
-     * @param[in]  lower_thr        Lower threshold.
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *magnitude, const ICLTensor *phase, ICLTensor *output, int32_t lower_thr, bool border_undefined);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-    BorderSize border_size() const override;
-
-private:
-    const ICLTensor *_magnitude; /**< Source tensor - Magnitude. */
-    const ICLTensor *_phase;     /**< Source tensor - Quantized phase. */
-    ICLTensor       *_output;    /**< Destination tensor. */
-};
-
-/** OpenCL kernel to perform Edge tracing.
- */
-class CLEdgeTraceKernel : public ICLKernel
-{
-public:
-    /** Constructor */
-    CLEdgeTraceKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLEdgeTraceKernel(const CLEdgeTraceKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLEdgeTraceKernel &operator=(const CLEdgeTraceKernel &) = delete;
-    /** Initialise the kernel's source, destination and border mode.
-     *
-     * @param[in]     input            Source tensor. Data types supported: U16/U32.
-     * @param[out]    output           Destination tensor. Data types supported: U8.
-     * @param[in]     upper_thr        Upper threshold used for the hysteresis
-     * @param[in]     lower_thr        Lower threshold used for the hysteresis
-     * @param[in,out] visited          Tensor for keeping the visited pixels. Data types supported: U32.
-     *                                 Expected to be initialized to 0 before each run.
-     * @param[in,out] recorded         Tensor for keeping the recorded pixels. Data types supported: U32
-     *                                 Expected to be initialized to 0 before each run.
-     * @param[in,out] l1_stack         Tensor with the L1 stack for each pixel. Data types supported: S32.
-     *                                 Expected to be initialized to 0 before each run.
-     * @param[in,out] l1_stack_counter Tensor for counting the elements in the L1 stack of each pixel. Data types supported: U8.
-     *                                              Expected to be initialized to 0 before each run.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, int32_t upper_thr, int32_t lower_thr,
-                   ICLTensor *visited, ICLTensor *recorded, ICLTensor *l1_stack, ICLTensor *l1_stack_counter);
-    /** Initialise the kernel's source, destination and border mode.
-     *
-     * @param[in]     compile_context  The compile context to be used.
-     * @param[in]     input            Source tensor. Data types supported: U16/U32.
-     * @param[out]    output           Destination tensor. Data types supported: U8.
-     * @param[in]     upper_thr        Upper threshold used for the hysteresis
-     * @param[in]     lower_thr        Lower threshold used for the hysteresis
-     * @param[in,out] visited          Tensor for keeping the visited pixels. Data types supported: U32.
-     *                                 Expected to be initialized to 0 before each run.
-     * @param[in,out] recorded         Tensor for keeping the recorded pixels. Data types supported: U32
-     *                                 Expected to be initialized to 0 before each run.
-     * @param[in,out] l1_stack         Tensor with the L1 stack for each pixel. Data types supported: S32.
-     *                                 Expected to be initialized to 0 before each run.
-     * @param[in,out] l1_stack_counter Tensor for counting the elements in the L1 stack of each pixel. Data types supported: U8.
-     *                                              Expected to be initialized to 0 before each run.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, int32_t upper_thr, int32_t lower_thr,
-                   ICLTensor *visited, ICLTensor *recorded, ICLTensor *l1_stack, ICLTensor *l1_stack_counter);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;            /**< Source tensor. */
-    ICLTensor       *_output;           /**< Destination tensor. */
-    int32_t          _lower_thr;        /**< Lower threshold used for the hysteresis. */
-    int32_t          _upper_thr;        /**< Upper threshold used for the hysteresis. */
-    ICLTensor       *_visited;          /**< Marks visited elements */
-    ICLTensor       *_recorded;         /**< Marks recorded elements */
-    ICLTensor       *_l1_stack;         /**< L1 hysteris stack */
-    ICLTensor       *_l1_stack_counter; /**< L1 hysteris stack counter */
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLCANNYEDGEKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLChannelCombineKernel.h b/arm_compute/core/CL/kernels/CLChannelCombineKernel.h
deleted file mode 100644
index f9c33df7c1..0000000000
--- a/arm_compute/core/CL/kernels/CLChannelCombineKernel.h
+++ /dev/null
@@ -1,102 +0,0 @@
-/*
- * Copyright (c) 2016-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLCHANNELCOMBINEKERNEL_H
-#define ARM_COMPUTE_CLCHANNELCOMBINEKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-#include <array>
-#include <cstdint>
-
-namespace arm_compute
-{
-class ICLMultiImage;
-class ICLTensor;
-using ICLImage = ICLTensor;
-
-/** Interface for the channel combine kernel */
-class CLChannelCombineKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLChannelCombineKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLChannelCombineKernel(const CLChannelCombineKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLChannelCombineKernel &operator=(const CLChannelCombineKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLChannelCombineKernel(CLChannelCombineKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLChannelCombineKernel &operator=(CLChannelCombineKernel &&) = default;
-    /** Default destructor */
-    ~CLChannelCombineKernel() = default;
-    /** Configure function's inputs and outputs.
-     *
-     * @param[in]  plane0 The 2D plane that forms channel 0. Must be of U8 format.
-     * @param[in]  plane1 The 2D plane that forms channel 1. Must be of U8 format.
-     * @param[in]  plane2 The 2D plane that forms channel 2. Must be of U8 format.
-     * @param[in]  plane3 The 2D plane that forms channel 3. Must be of U8 format.
-     * @param[out] output The single planar output tensor. Supported formats: RGB888/RGBA8888/YUYV422/UYVY422.
-     */
-    void configure(const ICLTensor *plane0, const ICLTensor *plane1, const ICLTensor *plane2, const ICLTensor *plane3, ICLTensor *output);
-    /** Configure function's inputs and outputs.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  plane0          The 2D plane that forms channel 0. Must be of U8 format.
-     * @param[in]  plane1          The 2D plane that forms channel 1. Must be of U8 format.
-     * @param[in]  plane2          The 2D plane that forms channel 2. Must be of U8 format.
-     * @param[in]  plane3          The 2D plane that forms channel 3. Must be of U8 format.
-     * @param[out] output          The single planar output tensor.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *plane0, const ICLTensor *plane1, const ICLTensor *plane2, const ICLTensor *plane3, ICLTensor *output);
-    /** Configure function's inputs and outputs.
-     *
-     * @param[in]  plane0 The 2D plane that forms channel 0. Must be of U8 format.
-     * @param[in]  plane1 The 2D plane that forms channel 1. Must be of U8 format.
-     * @param[in]  plane2 The 2D plane that forms channel 2. Must be of U8 format.
-     * @param[out] output The multi planar output tensor. Supported formats: RGB888/RGBA8888/YUYV422/UYVY422.
-     */
-    void configure(const ICLImage *plane0, const ICLImage *plane1, const ICLImage *plane2, ICLMultiImage *output);
-    /** Configure function's inputs and outputs.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  plane0          The 2D plane that forms channel 0. Must be of U8 format.
-     * @param[in]  plane1          The 2D plane that forms channel 1. Must be of U8 format.
-     * @param[in]  plane2          The 2D plane that forms channel 2. Must be of U8 format.
-     * @param[out] output          The multi planar output tensor. Supported formats: RGB888/RGBA8888/YUYV422/UYVY422.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLImage *plane0, const ICLImage *plane1, const ICLImage *plane2, ICLMultiImage *output);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    std::array<const ICLTensor *, 4> _planes;
-    ICLTensor     *_output;
-    ICLMultiImage *_output_multi;
-    std::array<uint32_t, 3> _x_subsampling;
-    std::array<uint32_t, 3> _y_subsampling;
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLCHANNELCOMBINEKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLChannelExtractKernel.h b/arm_compute/core/CL/kernels/CLChannelExtractKernel.h
deleted file mode 100644
index 1ccf38bb8c..0000000000
--- a/arm_compute/core/CL/kernels/CLChannelExtractKernel.h
+++ /dev/null
@@ -1,95 +0,0 @@
-/*
- * Copyright (c) 2016-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLCHANNELEXTRACTKERNEL_H
-#define ARM_COMPUTE_CLCHANNELEXTRACTKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-#include "arm_compute/core/Types.h"
-
-#include <cstdint>
-
-namespace arm_compute
-{
-class ICLMultiImage;
-class ICLTensor;
-using ICLImage = ICLTensor;
-
-/** Interface for the channel extract kernel */
-class CLChannelExtractKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLChannelExtractKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLChannelExtractKernel(const CLChannelExtractKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLChannelExtractKernel &operator=(const CLChannelExtractKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLChannelExtractKernel(CLChannelExtractKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLChannelExtractKernel &operator=(CLChannelExtractKernel &&) = default;
-    /** Default destructor */
-    ~CLChannelExtractKernel() = default;
-    /** Set the input and output of the kernel
-     *
-     * @param[in]  input   Source tensor. Formats supported: RGB888/RGBA8888/YUYV422/UYVY422
-     * @param[in]  channel Channel to extract.
-     * @param[out] output  Destination tensor. Must be of U8 format.
-     */
-    void configure(const ICLTensor *input, Channel channel, ICLTensor *output);
-    /** Set the input and output of the kernel
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Source tensor. Formats supported: RGB888/RGBA8888/YUYV422/UYVY422
-     * @param[in]  channel         Channel to extract.
-     * @param[out] output          Destination tensor. Must be of U8 format.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, Channel channel, ICLTensor *output);
-    /** Set the input and output of the kernel
-     *
-     * @param[in]  input   Multi-planar source image. Formats supported: NV12/NV21/IYUV/YUV444
-     * @param[in]  channel Channel to extract.
-     * @param[out] output  Single-planar 2D destination image. Must be of U8 format.
-     */
-    void configure(const ICLMultiImage *input, Channel channel, ICLImage *output);
-    /** Set the input and output of the kernel
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Multi-planar source image. Formats supported: NV12/NV21/IYUV/YUV444
-     * @param[in]  channel         Channel to extract.
-     * @param[out] output          Single-planar 2D destination image. Must be of U8 format.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLMultiImage *input, Channel channel, ICLImage *output);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;
-    ICLTensor       *_output;
-    uint32_t         _num_elems_processed_per_iteration;
-    uint32_t         _subsampling;
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLCHANNELEXTRACTKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLChannelShuffleLayerKernel.h b/arm_compute/core/CL/kernels/CLChannelShuffleLayerKernel.h
deleted file mode 100644
index bf58525248..0000000000
--- a/arm_compute/core/CL/kernels/CLChannelShuffleLayerKernel.h
+++ /dev/null
@@ -1,82 +0,0 @@
-/*
- * Copyright (c) 2018-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLCHANNELSHUFFLELAYERKERNEL_H
-#define ARM_COMPUTE_CLCHANNELSHUFFLELAYERKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the channel shuffle kernel */
-class CLChannelShuffleLayerKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLChannelShuffleLayerKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLChannelShuffleLayerKernel(const CLChannelShuffleLayerKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLChannelShuffleLayerKernel &operator=(const CLChannelShuffleLayerKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLChannelShuffleLayerKernel(CLChannelShuffleLayerKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLChannelShuffleLayerKernel &operator=(CLChannelShuffleLayerKernel &&) = default;
-    /** Default destructor */
-    ~CLChannelShuffleLayerKernel() = default;
-    /** Configure function's inputs and outputs.
-     *
-     * @param[in]  input      Input tensor. Data types supported: All.
-     * @param[out] output     Output tensor. Data type supported: Same as @p input
-     * @param[in]  num_groups Number of groups. Must be greater than 1 and the number of channels of the tensors must be a multiple of the number of groups.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, unsigned int num_groups);
-    /** Configure function's inputs and outputs.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Input tensor. Data types supported: All.
-     * @param[out] output          Output tensor. Data type supported: Same as @p input
-     * @param[in]  num_groups      Number of groups. Must be greater than 1 and the number of channels of the tensors must be a multiple of the number of groups.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, unsigned int num_groups);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLChannelShuffleLayerKernel
-     *
-     * @param[in] input      Input tensor info. Data types supported: All.
-     * @param[in] output     Output tensor info. Data type supported: Same as @p input
-     * @param[in] num_groups Number of groups. Must be greater than 1 and the number of channels of the tensors must be a multiple of the number of groups.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output, unsigned int num_groups);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;
-    ICLTensor       *_output;
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLCHANNELSHUFFLELAYERKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLCol2ImKernel.h b/arm_compute/core/CL/kernels/CLCol2ImKernel.h
deleted file mode 100644
index c3a1ff3a50..0000000000
--- a/arm_compute/core/CL/kernels/CLCol2ImKernel.h
+++ /dev/null
@@ -1,106 +0,0 @@
-/*
- * Copyright (c) 2017-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLCOL2IMKERNEL_H
-#define ARM_COMPUTE_CLCOL2IMKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the col2im reshaping kernel.
- *
- * Rearranges each matrix column into image blocks. It's the inverse operation of @ref CLIm2ColKernel.
- *
- * For example, a vector of 9 elements can be reshaped to a block(image) of 3x3:
- *
- * @f[
- * \left( \begin{array}{ccccccccc}
- * a0 & a1 & a2 & a3 & a4 & a5 & a6 & a7 & a8 \\
- * \end{array} \right)
- * \rightarrow
- * \left( \begin{array}{ccc}
- * a0 & a1 & a2 \\
- * a3 & a4 & a5 \\
- * a6 & a7 & a8 \\
- * \end{array} \right)
- * @f]
- */
-class CLCol2ImKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLCol2ImKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLCol2ImKernel(const CLCol2ImKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLCol2ImKernel &operator=(const CLCol2ImKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLCol2ImKernel(CLCol2ImKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLCol2ImKernel &operator=(CLCol2ImKernel &&) = default;
-    /** Default destructor */
-    ~CLCol2ImKernel() = default;
-    /** Set the input and output of the kernel.
-     *
-     * @param[in]  input          The input tensor to convert. Data types supported: QASYMM8/QASYMM8_SIGNED/F16/F32
-     * @param[out] output         The output tensor. 3 lower dimensions represent a single output [width, height, OFM],
-     *                            while the rest represent batch of outputs. Data types supported: Same as @p input. Data layout: NCHW
-     * @param[in]  convolved_dims Output convolved dimensions.
-     * @param[in]  num_groups     (Optional) Number of groups when performing a grouped convolution
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, const Size2D &convolved_dims, unsigned int num_groups = 1);
-    /** Set the input and output of the kernel.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           The input tensor to convert. Data types supported: QASYMM8/QASYMM8_SIGNED/F16/F32
-     * @param[out] output          The output tensor. 3 lower dimensions represent a single output [width, height, OFM],
-     *                             while the rest represent batch of outputs. Data types supported: Same as @p input. Data layout: NCHW
-     * @param[in]  convolved_dims  Output convolved dimensions.
-     * @param[in]  num_groups      (Optional) Number of groups when performing a grouped convolution
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, const Size2D &convolved_dims, unsigned int num_groups = 1);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLCol2ImKernel
-     *
-     * @param[in] input          The input tensor to convert. Data types supported: QASYMM8/QASYMM8_SIGNED/F16/F32
-     * @param[in] output         The output tensor. 3 lower dimensions represent a single output [width, height, OFM],
-     *                           while the rest represent batch of outputs. Data types supported: Same as @p input. Data layout: NCHW
-     * @param[in] convolved_dims Output convolved dimensions.
-     * @param[in] num_groups     (Optional) Number of groups when performing a grouped convolution
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output, const Size2D &convolved_dims, unsigned int num_groups = 1);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-public:
-    const ICLTensor *_input;
-    ICLTensor       *_output;
-    Size2D           _convolved_dims;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLCOL2IMKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLColorConvertKernel.h b/arm_compute/core/CL/kernels/CLColorConvertKernel.h
deleted file mode 100644
index d57bb3de03..0000000000
--- a/arm_compute/core/CL/kernels/CLColorConvertKernel.h
+++ /dev/null
@@ -1,121 +0,0 @@
-/*
- * Copyright (c) 2016-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLCOLORCONVERTKERNEL_H
-#define ARM_COMPUTE_CLCOLORCONVERTKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLMultiImage;
-class ICLTensor;
-using ICLImage = ICLTensor;
-
-/** Interface for the color convert kernel.
- *
- */
-class CLColorConvertKernel : public ICLKernel
-{
-public:
-    /** Default constructor. */
-    CLColorConvertKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLColorConvertKernel(const CLColorConvertKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLColorConvertKernel &operator=(const CLColorConvertKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLColorConvertKernel(CLColorConvertKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLColorConvertKernel &operator=(CLColorConvertKernel &&) = default;
-    /** Default destructor. */
-    ~CLColorConvertKernel() = default;
-
-    /** Set the input and output of the kernel
-     *
-     * @param[in]  input  Source tensor. Formats supported: RGBA8888/UYVY422/YUYV422/RGB888
-     * @param[out] output Destination tensor. Formats supported: RGB888 (if the formats of @p input are RGBA8888/UYVY422/YUYV422),
-     *                                                          RGBA8888 (if the formats of @p input are UYVY422/YUYV422/RGB888/),
-     *                                                          U8 (if the formats of @p input is RGB888)
-     */
-    void configure(const ICLTensor *input, ICLTensor *output);
-    /** Set the input and output of the kernel
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Source tensor. Formats supported: RGBA8888/UYVY422/YUYV422/RGB888
-     * @param[out] output          Destination tensor. Formats supported: RGB888 (if the formats of @p input are RGBA8888/UYVY422/YUYV422),
-     *                                                          RGBA8888 (if the formats of @p input are UYVY422/YUYV422/RGB888/),
-     *                                                          U8 (if the formats of @p input is RGB888)
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output);
-    /** Set the input and output of the kernel
-     *
-     * @param[in]  input  Multi-planar source image. Formats supported: NV12/NV21/IYUV
-     * @param[out] output Single-planar destination image. Formats supported: RGB888/RGBA8888
-     */
-    void configure(const ICLMultiImage *input, ICLImage *output);
-    /** Set the input and output of the kernel
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Multi-planar source image. Formats supported: NV12/NV21/IYUV
-     * @param[out] output          Single-planar destination image. Formats supported: RGB888/RGBA8888
-     */
-    void configure(const CLCompileContext &compile_context, const ICLMultiImage *input, ICLImage *output);
-    /** Set the input and output of the kernel
-     *
-     * @param[in]  input  Single-planar source image. Formats supported: RGB888/RGBA8888/UYVY422/YUYV422
-     * @param[out] output Multi-planar destination image. Formats supported: NV12/IYUV/YUV444 (if the formats of @p input are RGB888/RGB8888)
-     */
-    void configure(const ICLImage *input, ICLMultiImage *output);
-    /** Set the input and output of the kernel
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Single-planar source image. Formats supported: RGB888/RGBA8888/UYVY422/YUYV422
-     * @param[out] output          Multi-planar destination image. Formats supported: NV12/IYUV/YUV444 (if the formats of @p input are RGB888/RGB8888)
-     */
-    void configure(const CLCompileContext &compile_context, const ICLImage *input, ICLMultiImage *output);
-    /** Set the input and output of the kernel
-     *
-     * @param[in]  input  Multi-planar source image. Formats supported: NV12/NV21/IYUV
-     * @param[out] output Multi-planar destination image. Formats supported: YUV444/IYUV (if the formats of @p input are NV12/NV21)/NV12 (if the format of  @p input is IYUV)
-     */
-    void configure(const ICLMultiImage *input, ICLMultiImage *output);
-    /** Set the input and output of the kernel
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Multi-planar source image. Formats supported: NV12/NV21/IYUV
-     * @param[out] output          Multi-planar destination image. Formats supported: YUV444/IYUV (if the formats of @p input are NV12/NV21)/NV12 (if the format of  @p input is IYUV)
-     */
-    void configure(const CLCompileContext &compile_context, const ICLMultiImage *input, ICLMultiImage *output);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor     *_input;        /*pointer to single planar tensor input */
-    ICLTensor           *_output;       /*pointer to single planar tensor output */
-    const ICLMultiImage *_multi_input;  /*pointer to multi-planar input */
-    ICLMultiImage       *_multi_output; /*pointer to multi-planar output */
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLCOLORCONVERTKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLComparisonKernel.h b/arm_compute/core/CL/kernels/CLComparisonKernel.h
deleted file mode 100644
index bbf5f19e2f..0000000000
--- a/arm_compute/core/CL/kernels/CLComparisonKernel.h
+++ /dev/null
@@ -1,89 +0,0 @@
-/*
- * Copyright (c) 2018-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLCOMPARISONKERNEL_H
-#define ARM_COMPUTE_CLCOMPARISONKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-#include "arm_compute/core/Types.h"
-
-namespace arm_compute
-{
-// Forward declarations
-class ICLTensor;
-
-/** Interface for the comparison kernel. */
-class CLComparisonKernel : public ICLKernel
-{
-public:
-    /** Default constructor. */
-    CLComparisonKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLComparisonKernel(const CLComparisonKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLComparisonKernel &operator=(const CLComparisonKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLComparisonKernel(CLComparisonKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLComparisonKernel &operator=(CLComparisonKernel &&) = default;
-    /** Default destructor */
-    ~CLComparisonKernel() = default;
-    /** Set the inputs and output tensors
-     *
-     * @param[in]  input1    Source tensor. Data types supported: All.
-     * @param[in]  input2    Source tensor. Data types supported: Same as @p input1.
-     * @param[out] output    Destination tensor. Data types supported: U8.
-     * @param[in]  operation Comparison operation to use.
-     */
-    void configure(const ICLTensor *input1, const ICLTensor *input2, ICLTensor *output, ComparisonOperation operation);
-    /** Set the inputs and output tensors
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input1          Source tensor. Data types supported: All.
-     * @param[in]  input2          Source tensor. Data types supported: Same as @p input1.
-     * @param[out] output          Destination tensor. Data types supported: U8.
-     * @param[in]  operation       Comparison operation to use.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input1, const ICLTensor *input2, ICLTensor *output, ComparisonOperation operation);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLComparisonKernel
-     *
-     * @param[in] input1    Source tensor. Data types supported: All.
-     * @param[in] input2    Source tensor. Data types supported: Same as @p input1.
-     * @param[in] output    Destination tensor. Data types supported: U8.
-     * @param[in] operation Comparison operation to use.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input1, const ITensorInfo *input2, const ITensorInfo *output, ComparisonOperation operation);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-    BorderSize border_size() const override;
-
-private:
-    const ICLTensor *_input1; /**< Source tensor 1 */
-    const ICLTensor *_input2; /**< Source tensor 2 */
-    ICLTensor       *_output; /**< Destination tensor */
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLCOMPARISONKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLConvertFullyConnectedWeightsKernel.h b/arm_compute/core/CL/kernels/CLConvertFullyConnectedWeightsKernel.h
deleted file mode 100644
index 5d9e9bdd85..0000000000
--- a/arm_compute/core/CL/kernels/CLConvertFullyConnectedWeightsKernel.h
+++ /dev/null
@@ -1,90 +0,0 @@
-/*
- * Copyright (c) 2018-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLCONVERTFULLYCONNECTEDWEIGHTSKERNEL_H
-#define ARM_COMPUTE_CLCONVERTFULLYCONNECTEDWEIGHTSKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface to convert the 2D Fully Connected weights from NCHW to NHWC or vice versa.
- *
- * @note This function can be applied to the 2D weights used by a Fully Connected layer if:
- *       - It follows a Convolution layer
- *       - The data layout used by the network does not match the one the model has been trained in.
- *
- * @note This function assumes the weights are already reshaped (transposed)
- */
-class CLConvertFullyConnectedWeightsKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLConvertFullyConnectedWeightsKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLConvertFullyConnectedWeightsKernel(const CLConvertFullyConnectedWeightsKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLConvertFullyConnectedWeightsKernel &operator=(const CLConvertFullyConnectedWeightsKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLConvertFullyConnectedWeightsKernel(CLConvertFullyConnectedWeightsKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLConvertFullyConnectedWeightsKernel &operator=(CLConvertFullyConnectedWeightsKernel &&) = default;
-    /** Default destructor */
-    ~CLConvertFullyConnectedWeightsKernel() = default;
-    /** Set the input and output tensor.
-     *
-     * @param[in]  input                Source weights tensor to convert. Must be 2 dimensional. Data types supported: All.
-     * @param[out] output               The converted weights tensor. Shape and Data Type: Same as @p input.
-     * @param[in]  original_input_shape Shape of the original input tensor (the one entering fully connected layer).
-     * @param[in]  data_layout          The data layout the weights have been trained in.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, const TensorShape &original_input_shape, DataLayout data_layout);
-    /** Set the input and output tensor.
-     *
-     * @param[in]  compile_context      The compile context to be used.
-     * @param[in]  input                Source weights tensor to convert. Must be 2 dimensional. Data types supported: All.
-     * @param[out] output               The converted weights tensor. Shape and Data Type: Same as @p input.
-     * @param[in]  original_input_shape Shape of the original input tensor (the one entering fully connected layer).
-     * @param[in]  data_layout          The data layout the weights have been trained in.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, const TensorShape &original_input_shape, DataLayout data_layout);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLConvertFullyConnectedWeightsKernel
-     *
-     * @param[in] input                Source weights tensor info to convert. Must be 2 dimensional. Data types supported: All.
-     * @param[in] output               The converted weights tensor info. Shape and Data Type: Same as @p input.
-     * @param[in] original_input_shape Shape of the original input tensor (the one entering fully connected layer).
-     * @param[in] data_layout          The data layout the weights have been trained in.
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output, const TensorShape &original_input_shape, DataLayout data_layout);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;
-    ICLTensor       *_output;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLCONVERTFULLYCONNECTEDWEIGHTSKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLConvolutionKernel.h b/arm_compute/core/CL/kernels/CLConvolutionKernel.h
deleted file mode 100644
index 0f500fb63a..0000000000
--- a/arm_compute/core/CL/kernels/CLConvolutionKernel.h
+++ /dev/null
@@ -1,224 +0,0 @@
-/*
- * Copyright (c) 2016-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLCONVOLUTIONKERNEL_H
-#define ARM_COMPUTE_CLCONVOLUTIONKERNEL_H
-
-#include "arm_compute/core/CL/ICLSimple2DKernel.h"
-
-#include <cstdint>
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/****************************************************************************************\
- *                                    Square Convolution                                *
-\****************************************************************************************/
-
-/** Interface for the kernel to run an arbitrary size convolution on a tensor. (Currently supports 3x3, 5x5, 7x7 and 9x9).
- * The client can supply a convolution matrix \f$ C_{m,n} \f$.
- * @f{eqnarray}{
- *  k_0 &=& \frac{m}{2}  \\
- *  l_0 &=& \frac{n}{2}  \\
- *  sum &=& \sum_{k=0,l=0}^{k=m-1,l=n-1} input(x+k-k_0, y+l-l_0) C_{k,l}
- *  @f}
- *
- * @note The above equation for this function is similar to the default OpenCV Filter2D function,
- *       which actually computes a correlation and not a convolution.
- *       In case of a real convolution the convolution matrix should be flipped both horizontally and vertically.
- */
-template <unsigned int matrix_size>
-class CLConvolutionKernel : public ICLSimple2DKernel
-{
-public:
-    /** Initialise the kernel's input, output and border mode.
-     *
-     * @param[in]  input            Source tensor. Data types supported: U8.
-     * @param[out] output           Destination tensor, Data types supported: U8, S16.
-     * @param[in]  conv             Convolution matrix to apply to the input tensor.
-     * @param[in]  scale            Scale of the convolution matrix. If 0 is passed, it will be set to the sum of the coefficients of the convolution or 1 if they add up to 0.
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, const int16_t *conv, uint32_t scale, bool border_undefined);
-    /** Initialise the kernel's input, output and border mode.
-     *
-     * @param[in]  compile_context  The compile context to be used.
-     * @param[in]  input            Source tensor. Data types supported: U8.
-     * @param[out] output           Destination tensor, Data types supported: U8, S16.
-     * @param[in]  conv             Convolution matrix to apply to the input tensor.
-     * @param[in]  scale            Scale of the convolution matrix. If 0 is passed, it will be set to the sum of the coefficients of the convolution or 1 if they add up to 0.
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, const int16_t *conv, uint32_t scale, bool border_undefined);
-
-    // Inherited methods overridden:
-    BorderSize border_size() const override;
-};
-
-/** Interface for the kernel which applies a 3x3 convolution to a tensor. */
-using CLConvolution3x3Kernel = CLConvolutionKernel<3>;
-/** Interface for the kernel which applies a 5x5 convolution to a tensor. */
-using CLConvolution5x5Kernel = CLConvolutionKernel<5>;
-/** Interface for the kernel which applies a 7x7 convolution to a tensor. */
-using CLConvolution7x7Kernel = CLConvolutionKernel<7>;
-/** Interface for the kernel which applies a 9x9 convolution to a tensor. */
-using CLConvolution9x9Kernel = CLConvolutionKernel<9>;
-
-/****************************************************************************************\
- *                              Separable Square Convolution                            *
-\****************************************************************************************/
-
-/** Kernel for the Horizontal pass of a Separable Convolution. Currently support 5x5, 7x7, 9x9 */
-template <unsigned int matrix_size>
-class CLSeparableConvolutionHorKernel : public ICLSimple2DKernel
-{
-public:
-    /** Default Constructor */
-    CLSeparableConvolutionHorKernel();
-    /** Initialise the kernel's input, output and border mode.
-     *
-     * @param[in]  input            Source tensor. Data types supported: U8.
-     * @param[out] output           Destination tensor, Data types supported: S16.
-     * @param[in]  conv             Convolution matrix to apply to the input tensor.
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, const int16_t *conv, bool border_undefined);
-    /** Initialise the kernel's input, output and border mode.
-     *
-     * @param[in]  compile_context  The compile context to be used.
-     * @param[in]  input            Source tensor. Data types supported: U8.
-     * @param[out] output           Destination tensor, Data types supported: U16/S16/S32.
-     * @param[in]  conv             Convolution matrix to apply to the input tensor.
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, const int16_t *conv, bool border_undefined);
-
-    // Inherited methods overridden:
-    BorderSize border_size() const override;
-
-private:
-    BorderSize _border_size; /**< Border size */
-};
-
-/** Interface for the kernel which applies a horizontal pass of 5x5 convolution to a tensor. */
-using CLSeparableConvolution5x5HorKernel = CLSeparableConvolutionHorKernel<5>;
-/** Interface for the kernel which applies a horizontal pass of 7x7 convolution to a tensor. */
-using CLSeparableConvolution7x7HorKernel = CLSeparableConvolutionHorKernel<7>;
-/** Interface for the kernel which applies a horizontal pass of 9x9 convolution to a tensor. */
-using CLSeparableConvolution9x9HorKernel = CLSeparableConvolutionHorKernel<9>;
-
-/** Kernel for the Vertical pass of a Separable Convolution. Currently supports 5x5, 7x7, 9x9 */
-template <unsigned int matrix_size>
-class CLSeparableConvolutionVertKernel : public ICLSimple2DKernel
-{
-public:
-    /** Initialise the kernel's input, output and border mode.
-     *
-     * @param[in]  input            Source tensor. Data types supported: U16/S16/S32.
-     * @param[out] output           Destination tensor, Data types supported: U8, S16.
-     * @param[in]  conv             Convolution matrix to apply to the input tensor.
-     * @param[in]  scale            Scale of the convolution matrix.
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     * @param[in]  data_type        Data type to use for intermeidate result. @sa data_type_for_convolution
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, const int16_t *conv, uint32_t scale, bool border_undefined, DataType data_type = DataType::S32);
-    /** Initialise the kernel's input, output and border mode.
-     *
-     * @param[in]  compile_context  The compile context to be used.
-     * @param[in]  input            Source tensor. Data types supported: U16/S16/S32.
-     * @param[out] output           Destination tensor, Data types supported: U8, S16.
-     * @param[in]  conv             Convolution matrix to apply to the input tensor.
-     * @param[in]  scale            Scale of the convolution matrix.
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     * @param[in]  data_type        Data type to use for intermeidate result. @sa data_type_for_convolution
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, const int16_t *conv, uint32_t scale, bool border_undefined, DataType data_type = DataType::S32);
-
-    // Inherited methods overridden:
-    BorderSize border_size() const override;
-};
-
-/** Interface for the kernel which applies a vertical pass of 5x5 convolution to a tensor. */
-using CLSeparableConvolution5x5VertKernel = CLSeparableConvolutionVertKernel<5>;
-/** Interface for the kernel which applies a vertical pass of 7x7 convolution to a tensor. */
-using CLSeparableConvolution7x7VertKernel = CLSeparableConvolutionVertKernel<7>;
-/** Interface for the kernel which applies a vertical pass of 9x9 convolution to a tensor. */
-using CLSeparableConvolution9x9VertKernel = CLSeparableConvolutionVertKernel<9>;
-
-/****************************************************************************************\
- *                                 Rectangle Convolution                                *
-\****************************************************************************************/
-
-/** Kernel for the running convolution on a rectangle matrix.
- *
- * @note Supports combinations of 3,5,7 and 9.
- */
-class CLConvolutionRectangleKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLConvolutionRectangleKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLConvolutionRectangleKernel(const CLConvolutionRectangleKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLConvolutionRectangleKernel &operator=(const CLConvolutionRectangleKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLConvolutionRectangleKernel(CLConvolutionRectangleKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLConvolutionRectangleKernel &operator=(CLConvolutionRectangleKernel &&) = default;
-    /** Initialise the kernel's input, output and border mode.
-     *
-     * @param[in]  input            Source tensor. Data types supported: U8.
-     * @param[out] output           Destination tensor, Data types supported: U8, S16.
-     * @param[in]  conv             Convolution matrix to apply to the input tensor.
-     * @param[in]  width            Width of convolution matrix (Number of columns)
-     * @param[in]  height           Height of convolution matrix (Number of rows)
-     * @param[in]  scale            Scale of the convolution matrix. If 0 is passed, it will be set to the sum of the coefficients of the convolution or 1 if they add up to 0.
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, const int16_t *conv, uint32_t width, uint32_t height, uint32_t scale, bool border_undefined);
-    /** Initialise the kernel's input, output and border mode.
-     *
-     * @param[in]  compile_context  The compile context to be used.
-     * @param[in]  input            Source tensor. Data types supported: U8.
-     * @param[out] output           Destination tensor, Data types supported: U8, S16.
-     * @param[in]  conv             Convolution matrix to apply to the input tensor.
-     * @param[in]  width            Width of convolution matrix (Number of columns)
-     * @param[in]  height           Height of convolution matrix (Number of rows)
-     * @param[in]  scale            Scale of the convolution matrix. If 0 is passed, it will be set to the sum of the coefficients of the convolution or 1 if they add up to 0.
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, const int16_t *conv, uint32_t width, uint32_t height, uint32_t scale, bool border_undefined);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-    BorderSize border_size() const override;
-
-private:
-    BorderSize       _border_size;
-    const ICLTensor *_input;
-    ICLTensor       *_output;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLCONVOLUTIONKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLCopyKernel.h b/arm_compute/core/CL/kernels/CLCopyKernel.h
deleted file mode 100644
index 5c91e27935..0000000000
--- a/arm_compute/core/CL/kernels/CLCopyKernel.h
+++ /dev/null
@@ -1,83 +0,0 @@
-/*
- * Copyright (c) 2018-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLCOPYKERNEL_H
-#define ARM_COMPUTE_CLCOPYKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-#include "arm_compute/core/Types.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** OpenCL kernel to perform a copy between two tensors */
-class CLCopyKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLCopyKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers). */
-    CLCopyKernel(const CLCopyKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers). */
-    CLCopyKernel &operator=(const CLCopyKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLCopyKernel(CLCopyKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLCopyKernel &operator=(CLCopyKernel &&) = default;
-    /** Initialize the kernel's input, output.
-     *
-     * @param[in]  input         Source tensor. Data types supported: All.
-     * @param[out] output        Destination tensor. Data types supported: same as @p input.
-     * @param[in]  output_window (Optional) Window to be used in case only copying into part of a tensor. Default is nullptr.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, Window *output_window = nullptr);
-    /** Initialize the kernel's input, output.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Source tensor. Data types supported: All.
-     * @param[out] output          Destination tensor. Data types supported: same as @p input.
-     * @param[in]  output_window   (Optional) Window to be used in case only copying into part of a tensor. Default is nullptr.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, Window *output_window = nullptr);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLCopyKernel
-     *
-     * @param[in] input         Source tensor info. Data types supported: All.
-     * @param[in] output        Destination tensor info. Data types supported: same as @p input.
-     * @param[in] output_window (Optional) Window to be used in case only copying into part of a tensor. Default is nullptr.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output, Window *output_window = nullptr);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;
-    ICLTensor       *_output;
-    Window           _output_window;
-    bool             _has_output_window;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLCOPYKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLCropKernel.h b/arm_compute/core/CL/kernels/CLCropKernel.h
deleted file mode 100644
index 91d70e6c1b..0000000000
--- a/arm_compute/core/CL/kernels/CLCropKernel.h
+++ /dev/null
@@ -1,103 +0,0 @@
-/*
- * Copyright (c) 2019-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLCROPKERNEL_H
-#define ARM_COMPUTE_CLCROPKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-#include "arm_compute/core/Types.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** OpenCL kernel to perform a copy between two tensors */
-class CLCropKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLCropKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers). */
-    CLCropKernel(const CLCropKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers). */
-    CLCropKernel &operator=(const CLCropKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLCropKernel(CLCropKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLCropKernel &operator=(CLCropKernel &&) = default;
-    /** Configure kernel
-     *
-     * @note Supported tensor rank: up to 4
-     *
-     * @param[in]  input               Source tensor. Data type supported: All. Data layouts supported: NHWC.
-     * @param[out] output              Destination tensor. Data type supported: F32
-     * @param[in]  start               Coordinates of where to start cropping the image.
-     * @param[in]  end                 Coordinates of where to end cropping the image.
-     * @param[in]  batch_index         Fourth dimension index of the 3D image to crop in @p input.
-     * @param[in]  extrapolation_value Value to be used for values outside of the image. Default is 0.
-     * @param[in]  output_window       Output window to be used in case cropped image is being copied into a tensor. Default is nullptr.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, Coordinates2D start, Coordinates2D end, uint32_t batch_index, float extrapolation_value = 0, Window *output_window = nullptr);
-    /** Configure kernel
-     *
-     * @note Supported tensor rank: up to 4
-     *
-     * @param[in]  compile_context     The compile context to be used.
-     * @param[in]  input               Source tensor. Data type supported: All. Data layouts supported: NHWC.
-     * @param[out] output              Destination tensor. Data type supported: F32
-     * @param[in]  start               Coordinates of where to start cropping the image.
-     * @param[in]  end                 Coordinates of where to end cropping the image.
-     * @param[in]  batch_index         Fourth dimension index of the 3D image to crop in @p input.
-     * @param[in]  extrapolation_value Value to be used for values outside of the image. Default is 0.
-     * @param[in]  output_window       Output window to be used in case cropped image is being copied into a tensor. Default is nullptr.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, Coordinates2D start, Coordinates2D end, uint32_t batch_index, float extrapolation_value = 0,
-                   Window *output_window = nullptr);
-
-    /** Static function to check if given info will lead to a valid configuration of @ref CLStridedSliceKernel
-     *
-     * @note Supported tensor rank: up to 4
-     *
-     * @param[in] input               Source tensor info. Data type supported: All. Data layouts supported: NHWC.
-     * @param[in] output              Destination tensor info. Data type supported: F32
-     * @param[in] start               Coordinates of where to start cropping the image.
-     * @param[in] end                 Coordinates of where to end cropping the image.
-     * @param[in] batch_index         Fourth dimension index of the 3D image to crop in @p input.
-     * @param[in] extrapolation_value Value to be used for values outside of the image. Default is 0.
-     * @param[in] output_window       Output window to be used in case cropped image is being copied into a tensor. Default is nullptr.
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output, Coordinates2D start, Coordinates2D end, uint32_t batch_index, float extrapolation_value = 0,
-                           Window *output_window = nullptr);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;
-    ICLTensor       *_output;
-    Coordinates2D    _start;
-    uint32_t         _batch_index;
-    float            _extrapolation_value;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLCROPKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLDeconvolutionLayerUpsampleKernel.h b/arm_compute/core/CL/kernels/CLDeconvolutionLayerUpsampleKernel.h
deleted file mode 100644
index 84265a2329..0000000000
--- a/arm_compute/core/CL/kernels/CLDeconvolutionLayerUpsampleKernel.h
+++ /dev/null
@@ -1,86 +0,0 @@
-/*
- * Copyright (c) 2017-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLDECONVOLUTIONLAYERUPSAMPLEKERNEL_H
-#define ARM_COMPUTE_CLDECONVOLUTIONLAYERUPSAMPLEKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the Deconvolution layer kernel on OpenCL.
- */
-class CLDeconvolutionLayerUpsampleKernel : public ICLKernel
-{
-public:
-    /** Constructor */
-    CLDeconvolutionLayerUpsampleKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLDeconvolutionLayerUpsampleKernel(const CLDeconvolutionLayerUpsampleKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLDeconvolutionLayerUpsampleKernel &operator=(const CLDeconvolutionLayerUpsampleKernel &) = delete;
-    /** Default Move Constructor. */
-    CLDeconvolutionLayerUpsampleKernel(CLDeconvolutionLayerUpsampleKernel &&) = default;
-    /** Default move assignment operator */
-    CLDeconvolutionLayerUpsampleKernel &operator=(CLDeconvolutionLayerUpsampleKernel &&) = default;
-    /** Default destructor */
-    ~CLDeconvolutionLayerUpsampleKernel() = default;
-
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in]  input  Source tensor. Data types supported: All.
-     * @param[out] output Destination tensor. Data types supported: same as @p input. All but the lowest two dimensions must be the same size as in the input tensor, i.e. scaling is only performed within the XY-plane.
-     * @param[in]  info   Contains padding and stride information described in @ref PadStrideInfo.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, const PadStrideInfo &info);
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Source tensor. Data types supported: All.
-     * @param[out] output          Destination tensor. Data types supported: same as @p input. All but the lowest two dimensions must be the same size as in the input tensor, i.e. scaling is only performed within the XY-plane.
-     * @param[in]  info            Contains padding and stride information described in @ref PadStrideInfo.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, const PadStrideInfo &info);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLDeconvolutionLayerUpsample
-     *
-     * @param[in] input  Source tensor info. Data types supported: All.
-     * @param[in] output Destination tensor info. Data types supported: same as @p input. All but the lowest two dimensions must be the same size as in the input tensor, i.e. scaling is only performed within the XY-plane.
-     * @param[in] info   Contains padding and stride information described in @ref PadStrideInfo.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output, const PadStrideInfo &info);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;
-    ICLTensor       *_output;
-    PadStrideInfo    _info;
-    DataLayout       _data_layout;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLDECONVOLUTIONLAYERUPSAMPLEKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLDeconvolutionReshapeOutputKernel.h b/arm_compute/core/CL/kernels/CLDeconvolutionReshapeOutputKernel.h
deleted file mode 100644
index 688c943593..0000000000
--- a/arm_compute/core/CL/kernels/CLDeconvolutionReshapeOutputKernel.h
+++ /dev/null
@@ -1,106 +0,0 @@
-/*
- * Copyright (c) 2019-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLDECONVOLUTIONLAYERRESHAPEOUTPUTKERNEL_H
-#define ARM_COMPUTE_CLDECONVOLUTIONLAYERRESHAPEOUTPUTKERNEL_H
-
-#include "arm_compute/core/CL/ICLSimpleKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the OpenCL kernel to be used for reshaping the tensor before returning the result of deconvolution.
- *
- * The input tensor to this OpenCL kernel is expected to be the result of a @ref CLGEMM operation between the Deconvolution input and the Deconvolution filter.
- *
- * The input tensor should have the following shape: [filter_width * filter_height * ofms, width, height, batch_size]
- *
- * The output tensor should have the following shape: [stride_x * (input_width - 1) + filter_width - 2 * padx, stride_y * (input_height - 1) + filter_height - 2 * pady, ofms, batch_size]
- *
- * For example, given a tensor with dimensions [4, 2, 2] this function returns a tensor with dimensions [1, 4, 4].
- *
- */
-class CLDeconvolutionReshapeOutputKernel : public ICLSimpleKernel
-{
-public:
-    /** Default constructor */
-    CLDeconvolutionReshapeOutputKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLDeconvolutionReshapeOutputKernel(const CLDeconvolutionReshapeOutputKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLDeconvolutionReshapeOutputKernel &operator=(const CLDeconvolutionReshapeOutputKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLDeconvolutionReshapeOutputKernel(CLDeconvolutionReshapeOutputKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLDeconvolutionReshapeOutputKernel &operator=(CLDeconvolutionReshapeOutputKernel &&) = default;
-    /** Default destructor */
-    ~CLDeconvolutionReshapeOutputKernel() = default;
-
-    /** Initialise the kernel's source and destination.
-     *
-     * @param[in]  input        Input tensor. Supported data types: QASYMM8/QASYMM8_SIGNED/S32/F16/F32.
-     * @param[in]  bias         Bias tensor to be added directly during the reshape operation. Supported data types: same as @p input.  Supported data layouts: same as @p input.
-     * @param[out] output       Output tensor with the following shape: [stride_x * (input_width - 1) + filter_width - 2 * padx, stride_y * (input_height - 1) + filter_height - 2 * pady, ofms, batch_size]
-     *                          Supported data types: same as @p input.  Supported data layouts: same as @p input.
-     * @param[in]  input_info   Deconvolution input tensor info. Supported data types: same as @p input.  Supported data layouts: same as @p input.
-     * @param[in]  weights_info Deconvolution weights tensor info. Supported data types: same as @p input.  Supported data layouts: same as @p input.
-     * @param[in]  deconv_info  Contains padding and policies to be used in the deconvolution, this is described in @ref PadStrideInfo. This kernel supports only stride_x = weights.width && stride_y = weights.height. Moreover, padding is not supported.
-     */
-    void configure(const ICLTensor *input, const ICLTensor *bias, ICLTensor *output, const ITensorInfo *input_info, const ITensorInfo *weights_info, const PadStrideInfo &deconv_info);
-    /** Initialise the kernel's source and destination.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Input tensor. Supported data types: QASYMM8/QASYMM8_SIGNED/S32/F16/F32.
-     * @param[in]  bias            Bias tensor to be added directly during the reshape operation. Supported data types: same as @p input.  Supported data layouts: same as @p input.
-     * @param[out] output          Output tensor with the following shape: [stride_x * (input_width - 1) + filter_width - 2 * padx, stride_y * (input_height - 1) + filter_height - 2 * pady, ofms, batch_size]
-     *                             Supported data types: same as @p input.  Supported data layouts: same as @p input.
-     * @param[in]  input_info      Deconvolution input tensor info. Supported data types: same as @p input.  Supported data layouts: same as @p input.
-     * @param[in]  weights_info    Deconvolution weights tensor info. Supported data types: same as @p input.  Supported data layouts: same as @p input.
-     * @param[in]  deconv_info     Contains padding and policies to be used in the deconvolution, this is described in @ref PadStrideInfo. This kernel supports only stride_x = weights.width && stride_y = weights.height. Moreover, padding is not supported.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, const ICLTensor *bias, ICLTensor *output, const ITensorInfo *input_info, const ITensorInfo *weights_info,
-                   const PadStrideInfo &deconv_info);
-
-    /** Static function to check if given info will lead to a valid configuration of @ref  CLDeconvolutionReshapeOutputKernel.
-     *
-     * @param[in] input        GEMM output tensor info to be reshaped. Supported data types: QASYMM8/QASYMM8_SIGNED/S32/F16/F32.
-     * @param[in] bias         (Optional) Optional bias tensor info to be added directly during the reshape operation. Supported data types: same as @p input.  Supported data layouts: same as @p input.
-     * @param[in] output       Reshaped output tensor info. Supported data types: same as @p input.  Supported data layouts: same as @p input.
-     * @param[in] input_info   Original input tensor info. Supported data types: same as @p input.  Supported data layouts: same as @p input.
-     * @param[in] weights_info Original weights tensor info output. Supported data types: same as @p input.  Supported data layouts: same as @p input.
-     * @param[in] deconv_info  Contains padding and policies to be used in the deconvolution, this is described in @ref PadStrideInfo. This kernel supports only stride_x = weights.width && stride_y = weights.height. Moreover, padding is not supported.
-     *
-     * @return a Status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *bias, const ITensorInfo *output, const ITensorInfo *input_info, const ITensorInfo *weights_info, const PadStrideInfo &deconv_info);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    bool             _add_bias;
-    const ICLTensor *_bias;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLDECONVOLUTIONLAYERRESHAPEOUTPUTKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLDepthConcatenateLayerKernel.h b/arm_compute/core/CL/kernels/CLDepthConcatenateLayerKernel.h
deleted file mode 100644
index d8493bc5d8..0000000000
--- a/arm_compute/core/CL/kernels/CLDepthConcatenateLayerKernel.h
+++ /dev/null
@@ -1,80 +0,0 @@
-/*
- * Copyright (c) 2017-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-
-#ifndef ARM_COMPUTE_CLDEPTHCONCATENATEKERNEL_H
-#define ARM_COMPUTE_CLDEPTHCONCATENATEKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-#include "arm_compute/core/Types.h"
-
-namespace arm_compute
-{
-/** Interface for the depth concatenate kernel.
- *  The input tensor will be concatenated into the output tensor.
- */
-class CLDepthConcatenateLayerKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLDepthConcatenateLayerKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLDepthConcatenateLayerKernel(const CLDepthConcatenateLayerKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLDepthConcatenateLayerKernel &operator=(const CLDepthConcatenateLayerKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLDepthConcatenateLayerKernel(CLDepthConcatenateLayerKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLDepthConcatenateLayerKernel &operator=(CLDepthConcatenateLayerKernel &&) = default;
-    /** Default destructor */
-    ~CLDepthConcatenateLayerKernel() = default;
-    /** Initialise the kernel's inputs and output
-     *
-     * @param[in]     compile_context The compile context to be used.
-     * @param[in]     input           Input tensor. Data types supported: QASYMM8/QASYMM8_SIGNED/F16/F32.
-     * @param[in]     depth_offset    The offset on the Z axis.
-     * @param[in,out] output          Output tensor. Data types supported: Same as @p input.
-     *
-     * @note: The output tensor's low two dimensions can't be smaller than the input one's.
-     * @note: The gaps between the two lowest dimensions of input and output need to be divisible by 2.
-     *
-     */
-    void configure(const CLCompileContext &compile_context, ITensorInfo *input, unsigned int depth_offset, ITensorInfo *output);
-    /**  Static function to check if given info will lead to a valid configuration of @ref CLDepthConcatenateLayerKernel
-     *
-     * @param[in] input        Input tensor info. Data types supported: QASYMM8/QASYMM8_SIGNED/F16/F32
-     * @param[in] depth_offset The offset on the Z axis.
-     * @param[in] output       Output tensor info. Data types supported: Same as @p input.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, unsigned int depth_offset, const ITensorInfo *output);
-
-    // Inherited methods overridden:
-    void run_op(ITensorPack &tensors, const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    unsigned int _depth_offset;
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLDEPTHCONCATENATEKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLDepthConvertLayerKernel.h b/arm_compute/core/CL/kernels/CLDepthConvertLayerKernel.h
deleted file mode 100644
index 7f9696d835..0000000000
--- a/arm_compute/core/CL/kernels/CLDepthConvertLayerKernel.h
+++ /dev/null
@@ -1,91 +0,0 @@
-/*
- * Copyright (c) 2016-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLDEPTHCONVERTKERNEL_H
-#define ARM_COMPUTE_CLDEPTHCONVERTKERNEL_H
-
-#include "arm_compute/core/CL/ICLSimple3DKernel.h"
-#include "arm_compute/core/Types.h"
-
-#include <cstdint>
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the depth conversion kernel. */
-class CLDepthConvertLayerKernel : public ICLSimple3DKernel
-{
-public:
-    /** Set the input and output of the kernel.
-     *
-     * Valid conversions Input -> Output :
-     *
-     *   - QSYMM8_PER_CHANNEL -> QASYMM8 (ATTENTION: it is the user's responsibility to keep track of the quantization info in the TensorInfo meta-data)
-     *   - U8  -> S8, U16, S16, U32, S32, F16, F32
-     *   - U16 -> U8, S8, S16, U32, S32, F16, F32
-     *   - S16 -> U8, S8, U16, U32, S32, F16, F32
-     *   - U32 -> U8, S8, U16, S16, S32, F16, F32
-     *   - S32 -> U8, S8, U16, S16, U32, F16, F32
-     *   - F16 -> U8, S8, U16, S16, U32, F32
-     *   - F32 -> U8, S8, U16, S16, U32, F16
-     *
-     * @param[in]  input  The input tensor to convert. Data types supported: U8/S8/QSYMM8_PER_CHANNEL/U16/S16/U32/S32/F16/F32.
-     * @param[out] output The output tensor. Data types supported: U8/S8/QASYMM8/U16/S16/U32/S32/F16/F32.
-     * @param[in]  policy Conversion policy
-     * @param[in]  shift  Value for down/up conversions. Must be 0 <= shift < 8.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, ConvertPolicy policy, uint32_t shift);
-    /** Set the input and output of the kernel.
-     *
-     * Valid conversions Input -> Output :
-     *
-     *   - QSYMM8_PER_CHANNEL -> QASYMM8 (ATTENTION: it is the user's responsibility to keep track of the quantization info in the TensorInfo meta-data)
-     *   - U8  -> S8, U16, S16, U32, S32, F16, F32
-     *   - U16 -> U8, S8, S16, U32, S32, F16, F32
-     *   - S16 -> U8, S8, U16, U32, S32, F16, F32
-     *   - U32 -> U8, S8, U16, S16, S32, F16, F32
-     *   - S32 -> U8, S8, U16, S16, U32, F16, F32
-     *   - F16 -> U8, S8, U16, S16, U32, F32
-     *   - F32 -> U8, S8, U16, S16, U32, F16
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           The input tensor to convert. Data types supported: U8/S8/QSYMM8_PER_CHANNEL/U16/S16/U32/S32/F16/F32.
-     * @param[out] output          The output tensor. Data types supported: U8/S8/QASYMM8/U16/S16/U32/S32/F16/F32.
-     * @param[in]  policy          Conversion policy
-     * @param[in]  shift           Value for down/up conversions. Must be 0 <= shift < 8.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, ConvertPolicy policy, uint32_t shift);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLDepthConvertLayerKernel
-     *
-     * @param[in] input  Source tensor info. Data types supported: U8/S8/QSYMM8_PER_CHANNEL/U16/S16/U32/S32/F16/F32.
-     * @param[in] output Destination tensor info. Data type supported: U8/S8/QASYMM8/U16/S16/U32/S32/F16/F32.
-     * @param[in] policy Conversion policy
-     * @param[in] shift  Value for down/up conversions. Must be 0 <= shift < 8.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output, ConvertPolicy policy, uint32_t shift);
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLDEPTHCONVERTKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLDepthToSpaceLayerKernel.h b/arm_compute/core/CL/kernels/CLDepthToSpaceLayerKernel.h
deleted file mode 100644
index 1bd1e8e763..0000000000
--- a/arm_compute/core/CL/kernels/CLDepthToSpaceLayerKernel.h
+++ /dev/null
@@ -1,84 +0,0 @@
-/*
- * Copyright (c) 2019-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLDEPTHTOSPACELAYERKERNEL_H
-#define ARM_COMPUTE_CLDEPTHTOSPACELAYERKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-#include "arm_compute/core/Types.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the depth to space kernel */
-class CLDepthToSpaceLayerKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLDepthToSpaceLayerKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLDepthToSpaceLayerKernel(const CLDepthToSpaceLayerKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLDepthToSpaceLayerKernel &operator=(const CLDepthToSpaceLayerKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLDepthToSpaceLayerKernel(CLDepthToSpaceLayerKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLDepthToSpaceLayerKernel &operator=(CLDepthToSpaceLayerKernel &&) = default;
-    /** Default destructor */
-    ~CLDepthToSpaceLayerKernel() = default;
-    /** Initialise the kernel's inputs and output.
-     *
-     * @param[in]  input       Tensor input. Supported tensor rank: 4. Data types supported: All.
-     * @param[out] output      Tensor output. Data types supported: same as @p input
-     * @param[in]  block_shape Block shape value.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, int32_t block_shape);
-    /** Initialise the kernel's inputs and output.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Tensor input. Supported tensor rank: 4. Data types supported: All.
-     * @param[out] output          Tensor output. Data types supported: same as @p input
-     * @param[in]  block_shape     Block shape value.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, int32_t block_shape);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLDepthToSpaceLayerKernel.
-     *
-     * @param[in] input       Tensor input info. Supported tensor rank: 4. Data types supported: All.
-     * @param[in] output      Tensor output info. Data types supported: same as @p input
-     * @param[in] block_shape Block shape value.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output, int32_t block_shape);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;       /**< Source tensor */
-    ICLTensor       *_output;      /**< Destination tensor */
-    int32_t          _block_shape; /**< Block shape */
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLDEPTHTOSPACELAYERKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLDepthwiseConvolutionLayer3x3NCHWKernel.h b/arm_compute/core/CL/kernels/CLDepthwiseConvolutionLayer3x3NCHWKernel.h
deleted file mode 100644
index 93e7e374b0..0000000000
--- a/arm_compute/core/CL/kernels/CLDepthwiseConvolutionLayer3x3NCHWKernel.h
+++ /dev/null
@@ -1,114 +0,0 @@
-/*
- * Copyright (c) 2018-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLDEPTHWISECONVOLUTIONNCHWKERNEL3x3_H
-#define ARM_COMPUTE_CLDEPTHWISECONVOLUTIONNCHWKERNEL3x3_H
-
-#include "arm_compute/core/CL/kernels/ICLDepthwiseConvolutionLayer3x3Kernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the kernel to run a 3x3 depthwise convolution on a tensor when the data layout is NCHW.
- */
-class CLDepthwiseConvolutionLayer3x3NCHWKernel : public ICLDepthwiseConvolutionLayer3x3Kernel
-{
-public:
-    /** Default constructor */
-    CLDepthwiseConvolutionLayer3x3NCHWKernel();
-    /** Initialize the function's source, destination, conv and border_size.
-     *
-     * @param[in]  input              Source tensor. DataType supported: QASYMM8/QASYMM8_SIGNED/F16/F32.
-     * @param[in]  weights            Weights tensor. A 3D tensor with dimensions [3, 3, IFM].
-     *                                Data type supported: Same as @p input or QASYMM8/QASYMM8_SIGNED/QSYMM8_PER_CHANNEL when @p input is QASYMM8/QASYMM8_SIGNED.
-     * @param[in]  biases             Biases tensor. A 1D tensor with dimensions [IFM]. Must be nullptr if not needed.
-     *                                Data type supported: Same as @p input, S32 when input is QASYMM8/QASYMM8_SIGNED.
-     * @param[out] output             Destination tensor. Data type supported: Same as @p input.
-     * @param[in]  conv_info          Padding and stride information to use for the convolution.
-     * @param[in]  depth_multiplier   (Optional) Multiplier to apply to the input's depth in order to retrieve the output's depth. Defaults to 1.
-     * @param[in]  act_info           (Optional) Activation layer information in case of a fused activation. Only RELU, BOUNDED_RELU and LU_BOUNDED_RELU for QASYMM8 supported.
-     * @param[in]  dilation           (Optional) Dilation, in elements, across x and y. Defaults to (1, 1).
-     * @param[in]  output_multipliers (Optional) Output multipliers tensor for quantized computations. In case of per-channel quantization,
-     *                                the number of multipliers must be equal to the number of filters (IFM). Supported data types: S32
-     * @param[in]  output_shifts      (Optional) Output shifts tensor for quantized computations. In case of per-channel quantization,
-     *                                the number of multipliers must be equal to the number of filters (IFM). Supported data types: S32
-     */
-    void configure(const ICLTensor *input, const ICLTensor *weights, const ICLTensor *biases, ICLTensor *output, const PadStrideInfo &conv_info,
-                   unsigned int depth_multiplier = 1, ActivationLayerInfo act_info = ActivationLayerInfo(), const Size2D &dilation = Size2D(1U, 1U),
-                   const ICLTensor *output_multipliers = nullptr, const ICLTensor *output_shifts = nullptr) override;
-    /** Initialize the function's source, destination, conv and border_size.
-     *
-     * @param[in]  compile_context    The compile context to be used.
-     * @param[in]  input              Source tensor. DataType supported: QASYMM8/QASYMM8_SIGNED/F16/F32.
-     * @param[in]  weights            Weights tensor. A 3D tensor with dimensions [3, 3, IFM].
-     *                                Data type supported: Same as @p input or QASYMM8/QASYMM8_SIGNED/QSYMM8_PER_CHANNEL when @p input is QASYMM8/QASYMM8_SIGNED.
-     * @param[in]  biases             Biases tensor. A 1D tensor with dimensions [IFM]. Must be nullptr if not needed.
-     *                                Data type supported: Same as @p input, S32 when input is QASYMM8/QASYMM8_SIGNED.
-     * @param[out] output             Destination tensor. Data type supported: Same as @p input.
-     * @param[in]  conv_info          Padding and stride information to use for the convolution.
-     * @param[in]  depth_multiplier   (Optional) Multiplier to apply to the input's depth in order to retrieve the output's depth. Defaults to 1.
-     * @param[in]  act_info           (Optional) Activation layer information in case of a fused activation. Only RELU, BOUNDED_RELU and LU_BOUNDED_RELU for QASYMM8 supported.
-     * @param[in]  dilation           (Optional) Dilation, in elements, across x and y. Defaults to (1, 1).
-     * @param[in]  output_multipliers (Optional) Output multipliers tensor for quantized computations. In case of per-channel quantization,
-     *                                the number of multipliers must be equal to the number of filters (IFM). Supported data types: S32
-     * @param[in]  output_shifts      (Optional) Output shifts tensor for quantized computations. In case of per-channel quantization,
-     *                                the number of multipliers must be equal to the number of filters (IFM). Supported data types: S32
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, const ICLTensor *weights, const ICLTensor *biases, ICLTensor *output, const PadStrideInfo &conv_info,
-                   unsigned int depth_multiplier = 1, ActivationLayerInfo act_info = ActivationLayerInfo(), const Size2D &dilation = Size2D(1U, 1U),
-                   const ICLTensor *output_multipliers = nullptr, const ICLTensor *output_shifts = nullptr) override;
-    /** Static function to check if given info will lead to a valid configuration of @ref CLDepthwiseConvolutionLayer3x3NCHWKernel
-     *
-     * @param[in] input              Source tensor info. DataType supported: QASYMM8/QASYMM8_SIGNED/F16/F32.
-     * @param[in] weights            Weights tensor info. A 3D tensor with dimensions [3, 3, IFM].
-     *                               Data type supported: Same as @p input or QASYMM8/QASYMM8_SIGNED/QSYMM8_PER_CHANNEL when @p input is QASYMM8/QASYMM8_SIGNED.
-     * @param[in] biases             Biases tensor info. A 1D tensor with dimensions [IFM]. Must be nullptr if not needed.
-     *                               Data type supported: Same as @p input, S32 when input is QASYMM8/QASYMM8_SIGNED.
-     * @param[in] output             Destination tensor. Data type supported: Same as @p input.
-     * @param[in] conv_info          Padding and stride information to use for the convolution.
-     * @param[in] depth_multiplier   (Optional) Multiplier to apply to the input's depth in order to retrieve the output's depth. Defaults to 1.
-     * @param[in] act_info           (Optional) Activation layer information in case of a fused activation. Only RELU, BOUNDED_RELU and LU_BOUNDED_RELU are supported.
-     * @param[in] gpu_target         (Optional) GPU target to validate the kernel for. Defaults to midgard.
-     * @param[in] dilation           (Optional) Dilation, in elements, across x and y. Defaults to (1, 1).
-     * @param[in] output_multipliers (Optional) Output multipliers tensor info for quantized computations. In case of per-channel quantization,
-     *                               the number of multipliers must be equal to the number of filters (IFM). Supported data types: S32
-     * @param[in] output_shifts      (Optional) Output shifts tensor for quantized computations. In case of per-channel quantization,
-     *                               the number of multipliers must be equal to the number of filters (IFM). Supported data types: S32
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *weights, const ITensorInfo *biases, const ITensorInfo *output, const PadStrideInfo &conv_info,
-                           unsigned int depth_multiplier = 1, ActivationLayerInfo act_info = ActivationLayerInfo(), GPUTarget gpu_target = GPUTarget::MIDGARD,
-                           const Size2D &dilation = Size2D(1U, 1U), const ITensorInfo *output_multipliers = nullptr, const ITensorInfo *output_shifts = nullptr);
-
-    void run(const Window &window, cl::CommandQueue &queue) override;
-    BorderSize border_size() const override;
-
-private:
-    unsigned int _conv_stride_x;
-    unsigned int _conv_pad_top;
-    unsigned int _conv_pad_left;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLDEPTHWISECONVOLUTIONNCHWKERNEL3x3_H */
diff --git a/arm_compute/core/CL/kernels/CLDepthwiseConvolutionLayer3x3NHWCKernel.h b/arm_compute/core/CL/kernels/CLDepthwiseConvolutionLayer3x3NHWCKernel.h
deleted file mode 100644
index db57439de0..0000000000
--- a/arm_compute/core/CL/kernels/CLDepthwiseConvolutionLayer3x3NHWCKernel.h
+++ /dev/null
@@ -1,113 +0,0 @@
-/*
- * Copyright (c) 2018-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLDEPTHWISECONVOLUTIONNHWCKERNEL3x3_H
-#define ARM_COMPUTE_CLDEPTHWISECONVOLUTIONNHWCKERNEL3x3_H
-
-#include "arm_compute/core/CL/kernels/ICLDepthwiseConvolutionLayer3x3Kernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the kernel to run a 3x3 depthwise convolution on a tensor when the data layout is NHWC.
- */
-class CLDepthwiseConvolutionLayer3x3NHWCKernel : public ICLDepthwiseConvolutionLayer3x3Kernel
-{
-public:
-    /** Default constructor */
-    CLDepthwiseConvolutionLayer3x3NHWCKernel();
-    /** Default move assignment operator. */
-    /** Initialize the function's source, destination, conv and border_size.
-     *
-     * @param[in]  input              Source tensor. DataType supported: QASYMM8/QASYMM8_SIGNED/F16/F32.
-     * @param[in]  weights            Weights tensor. A 3D tensor with dimensions [IFM, 3, 3].
-     *                                Data type supported: Same as @p input or QASYMM8/QASYMM8_SIGNED/QSYMM8_PER_CHANNEL when @p input is QASYMM8/QASYMM8_SIGNED.
-     * @param[in]  biases             Biases tensor. A 1D tensor with dimensions [IFM]. Must be nullptr if not needed.
-     *                                Data type supported: Same as @p input, S32 when input is QASYMM8/QASYMM8_SIGNED.
-     * @param[out] output             Destination tensor. Data type supported: Same as @p input.
-     * @param[in]  conv_info          Padding and stride information to use for the convolution.
-     * @param[in]  depth_multiplier   (Optional) Multiplier to apply to the input's depth in order to retrieve the output's depth. Defaults to 1.
-     * @param[in]  act_info           (Optional) Activation layer information in case of a fused activation. Only RELU, BOUNDED_RELU and LU_BOUNDED_RELU are supported.
-     * @param[in]  dilation           (Optional) Dilation, in elements, across x and y. Defaults to (1, 1).
-     * @param[in]  output_multipliers (Optional) Output multipliers tensor for quantized computations. In case of per-channel quantization,
-     *                                the number of multipliers must be equal to the number of filters (IFM). Supported data types: S32
-     * @param[in]  output_shifts      (Optional) Output shifts tensor for quantized computations. In case of per-channel quantization,
-     *                                the number of multipliers must be equal to the number of filters (IFM). Supported data types: S32
-     */
-    void configure(const ICLTensor *input, const ICLTensor *weights, const ICLTensor *biases, ICLTensor *output, const PadStrideInfo &conv_info,
-                   unsigned int depth_multiplier = 1, ActivationLayerInfo act_info = ActivationLayerInfo(), const Size2D &dilation = Size2D(1U, 1U),
-                   const ICLTensor *output_multipliers = nullptr, const ICLTensor *output_shifts = nullptr) override;
-    /** Initialize the function's source, destination, conv and border_size.
-     *
-     * @param[in]  compile_context    The compile context to be used.
-     * @param[in]  input              Source tensor. DataType supported: QASYMM8/QASYMM8_SIGNED/F16/F32.
-     * @param[in]  weights            Weights tensor. A 3D tensor with dimensions [IFM, 3, 3].
-     *                                Data type supported: Same as @p input or QASYMM8/QASYMM8_SIGNED/QSYMM8_PER_CHANNEL when @p input is QASYMM8/QASYMM8_SIGNED.
-     * @param[in]  biases             Biases tensor. A 1D tensor with dimensions [IFM]. Must be nullptr if not needed.
-     *                                Data type supported: Same as @p input, S32 when input is QASYMM8/QASYMM8_SIGNED.
-     * @param[out] output             Destination tensor. Data type supported: Same as @p input.
-     * @param[in]  conv_info          Padding and stride information to use for the convolution.
-     * @param[in]  depth_multiplier   (Optional) Multiplier to apply to the input's depth in order to retrieve the output's depth. Defaults to 1.
-     * @param[in]  act_info           (Optional) Activation layer information in case of a fused activation. Only RELU, BOUNDED_RELU and LU_BOUNDED_RELU are supported.
-     * @param[in]  dilation           (Optional) Dilation, in elements, across x and y. Defaults to (1, 1).
-     * @param[in]  output_multipliers (Optional) Output multipliers tensor for quantized computations. In case of per-channel quantization,
-     *                                the number of multipliers must be equal to the number of filters (IFM). Supported data types: S32
-     * @param[in]  output_shifts      (Optional) Output shifts tensor for quantized computations. In case of per-channel quantization,
-     *                                the number of multipliers must be equal to the number of filters (IFM). Supported data types: S32
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, const ICLTensor *weights, const ICLTensor *biases, ICLTensor *output, const PadStrideInfo &conv_info,
-                   unsigned int depth_multiplier = 1, ActivationLayerInfo act_info = ActivationLayerInfo(), const Size2D &dilation = Size2D(1U, 1U),
-                   const ICLTensor *output_multipliers = nullptr, const ICLTensor *output_shifts = nullptr) override;
-    /** Static function to check if given info will lead to a valid configuration of @ref CLDepthwiseConvolutionLayer3x3NHWCKernel
-     *
-     * @param[in] input              Source tensor info. DataType supported: QASYMM8/QASYMM8_SIGNED/F16/F32.
-     * @param[in] weights            Weights tensor info. A 3D tensor with dimensions [IFM, 3, 3].
-     *                               Data type supported: Same as @p input or QASYMM8/QASYMM8_SIGNED/QSYMM8_PER_CHANNEL when @p input is QASYMM8/QASYMM8_SIGNED.
-     * @param[in] biases             Biases tensor info. A 1D tensor with dimensions [IFM]. Must be nullptr if not needed.
-     *                               Data type supported: Same as @p input, S32 when input is QASYMM8/QASYMM8_SIGNED.
-     * @param[in] output             Destination tensor info. Data type supported: Same as @p input.
-     * @param[in] conv_info          Padding and stride information to use for the convolution.
-     * @param[in] depth_multiplier   (Optional) Multiplier to apply to the input's depth in order to retrieve the output's depth. Defaults to 1.
-     * @param[in] act_info           (Optional) Activation layer information in case of a fused activation. Only RELU, BOUNDED_RELU and LU_BOUNDED_RELU are supported.
-     * @param[in] dilation           (Optional) Dilation, in elements, across x and y. Defaults to (1, 1).
-     * @param[in] output_multipliers (Optional) Output multipliers tensor info for quantized computations. In case of per-channel quantization,
-     *                               the number of multipliers must be equal to the number of filters (IFM). Supported data types: S32
-     * @param[in] output_shifts      (Optional) Output shifts tensor for quantized computations. In case of per-channel quantization,
-     *                               the number of multipliers must be equal to the number of filters (IFM). Supported data types: S32
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *weights, const ITensorInfo *biases, const ITensorInfo *output, const PadStrideInfo &conv_info,
-                           unsigned int depth_multiplier = 1, ActivationLayerInfo act_info = ActivationLayerInfo(), const Size2D &dilation = Size2D(1U, 1U),
-                           const ITensorInfo *output_multipliers = nullptr, const ITensorInfo *output_shifts = nullptr);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-    BorderSize border_size() const override;
-
-private:
-    unsigned int _num_planes_processed_per_iteration;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLDEPTHWISECONVOLUTIONNHWCKERNEL3x3_H */
diff --git a/arm_compute/core/CL/kernels/CLDepthwiseConvolutionLayerNativeKernel.h b/arm_compute/core/CL/kernels/CLDepthwiseConvolutionLayerNativeKernel.h
deleted file mode 100644
index 03a0106cc9..0000000000
--- a/arm_compute/core/CL/kernels/CLDepthwiseConvolutionLayerNativeKernel.h
+++ /dev/null
@@ -1,131 +0,0 @@
-/*
- * Copyright (c) 2019-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLDEPTHWISECONVOLUTIONLAYERNATIVEKERNEL_H
-#define ARM_COMPUTE_CLDEPTHWISECONVOLUTIONLAYERNATIVEKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-#include "arm_compute/core/KernelDescriptors.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the kernel to run a MxN depthwise convolution. M and N are respectively the rows and columns of the filter
-    This kernel assumes that tensor for the weights is NOT reshaped (Native version) */
-class CLDepthwiseConvolutionLayerNativeKernel : public ICLKernel
-{
-public:
-    /** Default Constructor */
-    CLDepthwiseConvolutionLayerNativeKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLDepthwiseConvolutionLayerNativeKernel(const CLDepthwiseConvolutionLayerNativeKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLDepthwiseConvolutionLayerNativeKernel &operator=(const CLDepthwiseConvolutionLayerNativeKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLDepthwiseConvolutionLayerNativeKernel(CLDepthwiseConvolutionLayerNativeKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLDepthwiseConvolutionLayerNativeKernel &operator=(CLDepthwiseConvolutionLayerNativeKernel &&) = default;
-    /** Initialize the function's source, destination and parameters
-     *
-     * @param[in]  input              Source tensor. Data type supported: QASYMM8/QASYMM8_SIGNED/FP32/FP16. Data layout supported: NHWC
-     * @param[in]  weights            Weights tensor. A 3D tensor with dimensions [IFM, N, M].
-     *                                Data type supported: Same as @p input or QASYMM8/QASYMM8_SIGNED/QSYMM8_PER_CHANNEL when @p input is QASYMM8.
-     * @param[in]  biases             Biases tensor. A 1D tensor with dimensions [IFM]. Must be nullptr if not needed.
-     *                                Data type supported: Same as @p input, S32 when input is QASYMM8/QASYMM8_SIGNED.
-     * @param[out] output             Destination tensor. Data type supported: Same as @p input.
-     * @param[in]  dwc_weights_info   Depthwise convolution layer weights info to retrieve the number of output elements processed by each thread
-     * @param[in]  dwc_info           Depthwise convolution layer info
-     * @param[in]  conv_info          Padding and stride information to use for the convolution.
-     * @param[in]  depth_multiplier   (Optional) Multiplier to apply to the input's depth in order to retrieve the output's depth. Defaults to 1.
-     * @param[in]  dilation           (Optional) Dilation, in elements, across x and y. Defaults to (1, 1).
-     * @param[in]  output_multipliers (Optional) Output multipliers tensor for quantized computations. In case of per-channel quantization,
-     *                                the number of multipliers must be equal to the number of filters (IFM). Supported data types: S32
-     * @param[in]  output_shifts      (Optional) Output shifts tensor for quantized computations. In case of per-channel quantization,
-     *                                the number of multipliers must be equal to the number of filters (IFM). Supported data types: S32
-     */
-    void configure(const ICLTensor *input, const ICLTensor *weights, const ICLTensor *biases, ICLTensor *output, const DWCWeightsKernelInfo &dwc_weights_info,
-                   const DWCKernelInfo &dwc_info, const PadStrideInfo &conv_info, unsigned int depth_multiplier = 1, const Size2D &dilation = Size2D(1U, 1U),
-                   const ICLTensor *output_multipliers = nullptr, const ICLTensor *output_shifts = nullptr);
-    /** Initialize the function's source, destination and parameters
-     *
-     * @param[in]  compile_context    The compile context to be used.
-     * @param[in]  input              Source tensor. Data type supported: QASYMM8/QASYMM8_SIGNED/FP32/FP16. Data layout supported: NHWC
-     * @param[in]  weights            Weights tensor. A 3D tensor with dimensions [IFM, N, M].
-     *                                Data type supported: Same as @p input or QASYMM8/QASYMM8_SIGNED/QSYMM8_PER_CHANNEL when @p input is QASYMM8.
-     * @param[in]  biases             Biases tensor. A 1D tensor with dimensions [IFM]. Must be nullptr if not needed.
-     *                                Data type supported: Same as @p input, S32 when input is QASYMM8/QASYMM8_SIGNED.
-     * @param[out] output             Destination tensor. Data type supported: Same as @p input.
-     * @param[in]  dwc_weights_info   Depthwise convolution layer weights info to retrieve the number of output elements processed by each thread
-     * @param[in]  dwc_info           Depthwise convolution layer info
-     * @param[in]  conv_info          Padding and stride information to use for the convolution.
-     * @param[in]  depth_multiplier   (Optional) Multiplier to apply to the input's depth in order to retrieve the output's depth. Defaults to 1.
-     * @param[in]  dilation           (Optional) Dilation, in elements, across x and y. Defaults to (1, 1).
-     * @param[in]  output_multipliers (Optional) Output multipliers tensor for quantized computations. In case of per-channel quantization,
-     *                                the number of multipliers must be equal to the number of filters (IFM). Supported data types: S32
-     * @param[in]  output_shifts      (Optional) Output shifts tensor for quantized computations. In case of per-channel quantization,
-     *                                the number of multipliers must be equal to the number of filters (IFM). Supported data types: S32
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, const ICLTensor *weights, const ICLTensor *biases, ICLTensor *output, const DWCWeightsKernelInfo &dwc_weights_info,
-                   const DWCKernelInfo &dwc_info, const PadStrideInfo &conv_info, unsigned int depth_multiplier = 1, const Size2D &dilation = Size2D(1U, 1U),
-                   const ICLTensor *output_multipliers = nullptr, const ICLTensor *output_shifts = nullptr);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLDepthwiseConvolutionLayerNativeKernel
-     *
-     * @param[in] input              Source tensor info. Data type supported: QASYMM8/QASYMM8_SIGNED/FP32/FP16. Data layout supported: NHWC
-     * @param[in] weights            Weights tensor info. A 3D tensor with dimensions [IFM, N, M].
-     *                               Data type supported: Same as @p input or QASYMM8/QASYMM8_SIGNED/QSYMM8_PER_CHANNEL when @p input is QASYMM8.
-     * @param[in] biases             Biases tensor info. A 1D tensor with dimensions [IFM]. Must be nullptr if not needed.
-     *                               Data type supported: Same as @p input, S32 when input is QASYMM8/QASYMM8_SIGNED.
-     * @param[in] output             Destination tensor info. Data type supported: Same as @p input.
-     * @param[in] dwc_weights_info   Depthwise convolution layer weights info to retrieve the number of output elements processed by each thread
-     * @param[in] dwc_info           Depthwise convolution layer info
-     * @param[in] conv_info          Padding and stride information to use for the convolution.
-     * @param[in] depth_multiplier   (Optional) Multiplier to apply to the input's depth in order to retrieve the output's depth. Defaults to 1.
-     * @param[in] dilation           (Optional) Dilation, in elements, across x and y. Defaults to (1, 1).
-     * @param[in] output_multipliers (Optional) Output multipliers tensor for quantized computations. In case of per-channel quantization,
-     *                               the number of multipliers must be equal to the number of filters (IFM). Supported data types: S32
-     * @param[in] output_shifts      (Optional) Output shifts tensor for quantized computations. In case of per-channel quantization,
-     *                               the number of multipliers must be equal to the number of filters (IFM). Supported data types: S32
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *weights, const ITensorInfo *biases, const ITensorInfo *output, const DWCWeightsKernelInfo &dwc_weights_info,
-                           const DWCKernelInfo &dwc_info, const PadStrideInfo &conv_info, unsigned int depth_multiplier = 1, const Size2D &dilation = Size2D(1U, 1U),
-                           const ITensorInfo *output_multipliers = nullptr, const ITensorInfo *output_shifts = nullptr);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;
-    const ICLTensor *_weights;
-    const ICLTensor *_biases;
-    ICLTensor       *_output;
-    unsigned int     _depth_multiplier;
-    const ICLTensor *_output_multipliers;
-    const ICLTensor *_output_shifts;
-    bool             _is_quantized;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLDEPTHWISECONVOLUTIONLAYERNATIVEKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLDepthwiseConvolutionLayerReshapeWeightsKernel.h b/arm_compute/core/CL/kernels/CLDepthwiseConvolutionLayerReshapeWeightsKernel.h
deleted file mode 100644
index 51aaf17600..0000000000
--- a/arm_compute/core/CL/kernels/CLDepthwiseConvolutionLayerReshapeWeightsKernel.h
+++ /dev/null
@@ -1,85 +0,0 @@
-/*
- * Copyright (c) 2019-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLDEPTHWISECONVOLUTIONLAYERRESHAPEWEIGHTSKERNEL_H
-#define ARM_COMPUTE_CLDEPTHWISECONVOLUTIONLAYERRESHAPEWEIGHTSKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the kernel to reshape the weights of depthwise convolution. */
-class CLDepthwiseConvolutionLayerReshapeWeightsKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLDepthwiseConvolutionLayerReshapeWeightsKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLDepthwiseConvolutionLayerReshapeWeightsKernel(const CLDepthwiseConvolutionLayerReshapeWeightsKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLDepthwiseConvolutionLayerReshapeWeightsKernel &operator=(const CLDepthwiseConvolutionLayerReshapeWeightsKernel &) = delete;
-    /** Default Move Constructor. */
-    CLDepthwiseConvolutionLayerReshapeWeightsKernel(CLDepthwiseConvolutionLayerReshapeWeightsKernel &&) = default;
-    /** Default move assignment operator */
-    CLDepthwiseConvolutionLayerReshapeWeightsKernel &operator=(CLDepthwiseConvolutionLayerReshapeWeightsKernel &&) = default;
-
-    /** Initialize the function's source and destination.
-     *
-     * @param[in]  input  The input tensor of dimension [IFM, W, H]. Data types supported: All. Data layouts supported: NHWC
-     * @param[out] output The output tensor of dimension [W*H*C0, ceil(IFM/C0)]. C0 is the number of channels read by each thread. Data types supported: same as @p weights.
-     * @param[in]  info   Depthwise convolution information to reshape the input tensor.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, const DepthwiseConvolutionReshapeInfo &info);
-    /** Initialize the function's source and destination.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           The input tensor of dimension [IFM, W, H]. Data types supported: All. Data layouts supported: NHWC
-     * @param[out] output          The output tensor of dimension [W*H*C0, ceil(IFM/C0)]. C0 is the number of channels read by each thread. Data types supported: same as @p weights.
-     * @param[in]  info            Depthwise convolution information to reshape the input tensor.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, const DepthwiseConvolutionReshapeInfo &info);
-
-    /** Static function to check if given info will lead to a valid configuration of @ref CLDepthwiseConvolutionLayer3x3NHWCKernel
-     *
-     * @param[in] input  The input tensor info of dimension [IFM, W, H]. Data types supported: All. Data layouts supported: NHWC
-     * @param[in] output The output tensor info of dimension [W*H*C0, ceil(IFM/C0)]. C0 is the number of channels read by each thread. Data types supported: same as @p weights.
-     * @param[in] info   Depthwise convolution information to reshape the input tensor.
-     *
-     * @return a Status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output, const DepthwiseConvolutionReshapeInfo &info);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;
-    ICLTensor       *_output;
-
-    void configure_dot_product(const DepthwiseConvolutionReshapeInfo &info);
-    void configure_generic(const DepthwiseConvolutionReshapeInfo &info);
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLDEPTHWISECONVOLUTIONLAYERRESHAPEWEIGHTSKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLDequantizationLayerKernel.h b/arm_compute/core/CL/kernels/CLDequantizationLayerKernel.h
deleted file mode 100644
index 7a582da132..0000000000
--- a/arm_compute/core/CL/kernels/CLDequantizationLayerKernel.h
+++ /dev/null
@@ -1,79 +0,0 @@
-/*
- * Copyright (c) 2017-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLDEQUANTIZATIONLAYERKERNEL_H
-#define ARM_COMPUTE_CLDEQUANTIZATIONLAYERKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the dequantization layer kernel. */
-class CLDequantizationLayerKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLDequantizationLayerKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLDequantizationLayerKernel(const CLDequantizationLayerKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLDequantizationLayerKernel &operator=(const CLDequantizationLayerKernel &) = delete;
-    /** Default Move Constructor. */
-    CLDequantizationLayerKernel(CLDequantizationLayerKernel &&) = default;
-    /** Default move assignment operator */
-    CLDequantizationLayerKernel &operator=(CLDequantizationLayerKernel &&) = default;
-    /** Default destructor */
-    ~CLDequantizationLayerKernel() = default;
-    /** Set the input, output, min and max.
-     *
-     * @param[in]  input  Source tensor. Data types supported: QASYMM8/QASYMM8_SIGNED/QSYMM8_PER_CHANNEL/QSYMM8/QSYMM16.
-     * @param[out] output Destination tensor. Data types supported: F16/F32.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output);
-    /** Set the input, output, min and max.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Source tensor. Data types supported: QASYMM8/QASYMM8_SIGNED/QSYMM8_PER_CHANNEL/QSYMM8/QSYMM16.
-     * @param[out] output          Destination tensor. Data types supported: F16/F32.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLDequantizationLayerKernel
-     *
-     * @param[in] input  Input tensor info. Data types supported: QASYMM8/QASYMM8_SIGNED/QSYMM8_PER_CHANNEL/QSYMM8/QSYMM16.
-     * @param[in] output Output tensor info. Data types supported: F16/F32.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;
-    ICLTensor       *_output;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLDEQUANTIZATIONLAYERKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLDerivativeKernel.h b/arm_compute/core/CL/kernels/CLDerivativeKernel.h
deleted file mode 100644
index b49905a5e6..0000000000
--- a/arm_compute/core/CL/kernels/CLDerivativeKernel.h
+++ /dev/null
@@ -1,83 +0,0 @@
-/*
- * Copyright (c) 2016-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLDERIVATIVEKERNEL_H
-#define ARM_COMPUTE_CLDERIVATIVEKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the derivative kernel. */
-class CLDerivativeKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLDerivativeKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLDerivativeKernel(const CLDerivativeKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLDerivativeKernel &operator=(const CLDerivativeKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLDerivativeKernel(CLDerivativeKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLDerivativeKernel &operator=(CLDerivativeKernel &&) = default;
-    /** Default destructor */
-    ~CLDerivativeKernel() = default;
-    /** Initialise the kernel's sources, destination and border
-     *
-     * @note At least one of output_x or output_y must be set
-     *
-     * @param[in]  input            Source tensor. Data types supported: U8.
-     * @param[out] output_x         (Optional) Destination tensor for the X gradient, Data types supported: S16.
-     * @param[out] output_y         (Optional) Destination tensor for the Y gradient, Data types supported: S16.
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output_x, ICLTensor *output_y, bool border_undefined);
-    /** Initialise the kernel's sources, destination and border
-     *
-     * @note At least one of output_x or output_y must be set
-     *
-     * @param[in]  compile_context  The compile context to be used.
-     * @param[in]  input            Source tensor. Data types supported: U8.
-     * @param[out] output_x         (Optional) Destination tensor for the X gradient, Data types supported: S16.
-     * @param[out] output_y         (Optional) Destination tensor for the Y gradient, Data types supported: S16.
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output_x, ICLTensor *output_y, bool border_undefined);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-    BorderSize border_size() const override;
-
-private:
-    const ICLTensor *_input;            /**< Input tensor */
-    ICLTensor       *_output_x;         /**< Output tensor - Derivate along the X direction */
-    ICLTensor       *_output_y;         /**< Output tensor - Derivate along the Y direction */
-    bool             _run_derivative_x; /**< Do we need to run Derivative X ? */
-    bool             _run_derivative_y; /**< Do we need to run Derivative Y ? */
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLDERIVATIVEKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLDilateKernel.h b/arm_compute/core/CL/kernels/CLDilateKernel.h
deleted file mode 100644
index 747f8fa5ca..0000000000
--- a/arm_compute/core/CL/kernels/CLDilateKernel.h
+++ /dev/null
@@ -1,59 +0,0 @@
-/*
- * Copyright (c) 2016-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLDILATEKERNEL_H
-#define ARM_COMPUTE_CLDILATEKERNEL_H
-
-#include "arm_compute/core/CL/ICLSimple2DKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the dilate kernel.
- *
- */
-class CLDilateKernel : public ICLSimple2DKernel
-{
-public:
-    /**Initialise the kernel's input and output.
-     *
-     * @param[in]  input            An input tensor. Data types supported: U8
-     * @param[out] output           The output tensor. Data types supported: U8.
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, bool border_undefined);
-    /**Initialise the kernel's input and output.
-     *
-     * @param[in]  compile_context  The compile context to be used.
-     * @param[in]  input            An input tensor. Data types supported: U8
-     * @param[out] output           The output tensor. Data types supported: U8.
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, bool border_undefined);
-
-    // Inherited methods overridden:
-    BorderSize border_size() const override;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLDILATEKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLDirectConvolutionLayerKernel.h b/arm_compute/core/CL/kernels/CLDirectConvolutionLayerKernel.h
deleted file mode 100644
index 5281a0c306..0000000000
--- a/arm_compute/core/CL/kernels/CLDirectConvolutionLayerKernel.h
+++ /dev/null
@@ -1,125 +0,0 @@
-/*
- * Copyright (c) 2017-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLDIRECTCONVOLUTIONLAYERKERNEL_H
-#define ARM_COMPUTE_CLDIRECTCONVOLUTIONLAYERKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-#include "arm_compute/core/Types.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the  direct convolution kernel.
- */
-class CLDirectConvolutionLayerKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLDirectConvolutionLayerKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLDirectConvolutionLayerKernel(const CLDirectConvolutionLayerKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLDirectConvolutionLayerKernel &operator=(const CLDirectConvolutionLayerKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLDirectConvolutionLayerKernel(CLDirectConvolutionLayerKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLDirectConvolutionLayerKernel &operator=(CLDirectConvolutionLayerKernel &&) = default;
-    /** Default destructor */
-    ~CLDirectConvolutionLayerKernel() = default;
-    /** Set the input, weights, biases and output tensors.
-     *
-     * @note: DirectConvolution only works in the following configurations:
-     *        1x1 convolution with stride_x = 1/2/3, stride_y = 1/2/3
-     *        3x3 convolution with stride_x = 1/2, stride_y = 1/2
-     *        5x5 convolution with stride_x = 1/2, stride_y = 1/2
-     *        9x9 convolution with stride_x = 1/2, stride_y = 1/2
-     *
-     * @param[in]  input     The input tensor to convolve. 3 lower dimensions represent a single input [width, height, IFM],
-     *                       while every optional dimension from 4 and above represent a batch of inputs. Data types supported: QASYMM8_SIGNED/QASYMM8/F16/F32.
-     * @param[in]  weights   Weights tensor. Weights are 4D tensor with dimensions [kernel_x, kernel_y, IFM, OFM].
-     *                       The 3rd dimension must be the same as the input's volume 3rd dimension.
-     *                       Data type supported:Same as @p input.
-     * @param[in]  biases    Biases tensor. Biases are 1D tensor with dimension [OFM].
-     *                       Data type supported: Should match @p input data type, except for input of QASYMM8 and QASYMM8_SIGNED type where biases should be of S32 type
-     * @param[out] output    Output tensor.
-     *                       The 3rd dimensions must be equal to the 4th dimension of the @p kernels tensor. Data types supported: Same as @p input.
-     * @param[in]  conv_info Contains padding and stride information described in @ref PadStrideInfo.
-     */
-    void configure(const ICLTensor *input, const ICLTensor *weights, const ICLTensor *biases, ICLTensor *output, const PadStrideInfo &conv_info);
-    /** Set the input, weights, biases and output tensors.
-     *
-     * @note: DirectConvolution only works in the following configurations:
-     *        1x1 convolution with stride_x = 1/2/3, stride_y = 1/2/3
-     *        3x3 convolution with stride_x = 1/2, stride_y = 1/2
-     *        5x5 convolution with stride_x = 1/2, stride_y = 1/2
-     *        9x9 convolution with stride_x = 1/2, stride_y = 1/2
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           The input tensor to convolve. 3 lower dimensions represent a single input [width, height, IFM],
-     *                             while every optional dimension from 4 and above represent a batch of inputs. Data types supported: QASYMM8_SIGNED/QASYMM8/F16/F32.
-     * @param[in]  weights         Weights tensor. Weights are 4D tensor with dimensions [kernel_x, kernel_y, IFM, OFM].
-     *                             The 3rd dimension must be the same as the input's volume 3rd dimension.
-     *                             Data type supported:Same as @p input.
-     * @param[in]  biases          Biases tensor. Biases are 1D tensor with dimension [OFM].
-     *                             Data type supported: Should match @p input data type, except for input of QASYMM8 and QASYMM8_SIGNED type where biases should be of S32 type
-     * @param[out] output          Output tensor.
-     *                             The 3rd dimensions must be equal to the 4th dimension of the @p kernels tensor. Data types supported: Same as @p input.
-     * @param[in]  conv_info       Contains padding and stride information described in @ref PadStrideInfo.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, const ICLTensor *weights, const ICLTensor *biases, ICLTensor *output, const PadStrideInfo &conv_info);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLDirectConvolutionLayerKernel
-     *
-     * @param[in] input     The input tensor to convolve. 3 lower dimensions represent a single input [width, height, IFM],
-     *                      while every optional dimension from 4 and above represent a batch of inputs. Data types supported: QASYMM8_SIGNED/QASYMM8/F16/F32.
-     * @param[in] weights   Weights tensor. Weights are 4D tensor with dimensions [kernel_x, kernel_y, IFM, OFM].
-     *                      The 3rd dimension must be the same as the input's volume 3rd dimension.
-     *                      Data type supported:Same as @p input.
-     * @param[in] biases    Biases tensor. Biases are 1D tensor with dimension [OFM].
-     *                      Data type supported: Should match @p input data type, except for input of QASYMM8 and QASYMM8_SIGNED type where biases should be of S32 type.
-     * @param[in] output    Output tensor.
-     *                      The 3rd dimensions must be equal to the 4th dimension of the @p kernels tensor. Data types supported: Same as @p input.
-     * @param[in] conv_info Contains padding and stride information described in @ref PadStrideInfo.
-     * @param[in] target    Target GPU architecture.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *weights, const ITensorInfo *biases, const ITensorInfo *output, const PadStrideInfo &conv_info, const GPUTarget target);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-    BorderSize border_size() const override;
-
-public:
-    const ICLTensor *_input;
-    const ICLTensor *_biases;
-    const ICLTensor *_weights;
-    ICLTensor       *_output;
-    DataLayout       _data_layout;
-    BorderSize       _border_size;
-    int              _conv_stride_x;
-    int              _conv_stride_y;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLDIRECTCONVOLUTIONLAYERKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLElementWiseUnaryLayerKernel.h b/arm_compute/core/CL/kernels/CLElementWiseUnaryLayerKernel.h
deleted file mode 100644
index 82cd953b68..0000000000
--- a/arm_compute/core/CL/kernels/CLElementWiseUnaryLayerKernel.h
+++ /dev/null
@@ -1,66 +0,0 @@
-/*
- * Copyright (c) 2018-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLELEMENTWISEUNARYLAYERKERNEL_H
-#define ARM_COMPUTE_CLELEMENTWISEUNARYLAYERKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-#include "arm_compute/core/CL/ICLSimpleKernel.h"
-#include "arm_compute/core/Types.h"
-
-namespace arm_compute
-{
-/** Interface for the elementwise unary operator */
-class CLElementWiseUnaryLayerKernel : public ICLKernel
-{
-public:
-    /** Initialise the kernel's inputs, output.
-     *
-     * @param[in]  input  First tensor input info. Data types supported: F16/F32.
-     * @param[out] output Output tensor info. Data types supported: Same as @p input.
-     * @param[in]  op     Element wise unary operation to perform.
-     */
-    void configure(const ITensorInfo *input, ITensorInfo *output, const ElementWiseUnary &op);
-    /** Initialise the kernel's inputs, output.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           First tensor input info. Data types supported: F16/F32.
-     * @param[out] output          Output tensor info. Data types supported: Same as @p input.
-     * @param[in]  op              Element wise unary operation to perform.
-     */
-    void configure(const CLCompileContext &compile_context, const ITensorInfo *input, ITensorInfo *output, const ElementWiseUnary &op);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLElementWiseUnaryLayerKernel
-     *
-     * @param[in] input  First tensor input info. Data types supported: F16/F32.
-     * @param[in] output Output tensor info. Data types supported: Same as @p input.
-     * @param[in] op     Element wise unary operation to perform.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output, const ElementWiseUnary &op);
-
-    // Inherited methods overridden:
-    void run_op(ITensorPack &tensors, const Window &window, cl::CommandQueue &queue) override;
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLELEMENTWISEUNARYLAYERKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLElementwiseOperationKernel.h b/arm_compute/core/CL/kernels/CLElementwiseOperationKernel.h
deleted file mode 100644
index 80737cb8eb..0000000000
--- a/arm_compute/core/CL/kernels/CLElementwiseOperationKernel.h
+++ /dev/null
@@ -1,212 +0,0 @@
-/*
- * Copyright (c) 2018-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLELEMENTWISEOPERATIONKERNEL_H
-#define ARM_COMPUTE_CLELEMENTWISEOPERATIONKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-#include "arm_compute/core/Types.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for an element-wise operation kernel
- *
- * Element-wise operation is computed by:
- * @f[ output(x,y) = OP(input1(x,y), input2(x,y))@f]
- *
- */
-class CLElementwiseOperationKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLElementwiseOperationKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLElementwiseOperationKernel(const CLElementwiseOperationKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLElementwiseOperationKernel &operator=(const CLElementwiseOperationKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLElementwiseOperationKernel(CLElementwiseOperationKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLElementwiseOperationKernel &operator=(CLElementwiseOperationKernel &&) = default;
-    /** Default destructor */
-    ~CLElementwiseOperationKernel() = default;
-
-    // Inherited methods overridden:
-    void run_op(ITensorPack &tensors, const Window &window, cl::CommandQueue &queue) override;
-
-protected:
-    /** The name of the operation */
-    virtual std::string name() = 0;
-
-    /** Initialise the kernel's output.
-     *
-     * @param[in] input1 First tensor input info. Data types supported: U8/S8/QASYMM8/QASYMM8_SIGNED/U16/S16/F16/U32/S32/F32.
-     * @param[in] input2 Second tensor input info. Data types supported: Same as @p input1.
-     * @param[in] output Output tensor info. Data types supported: Same as @p input1.
-     *
-     * @return a pair of Status and Window
-     */
-    virtual std::pair<Status, Window> validate_and_configure_window(ITensorInfo &input1, ITensorInfo &input2, ITensorInfo &output) = 0;
-
-    /** Generate the build options for the specific kernel
-     *
-     * @reutrn a CLBuildOptions struct
-     */
-    virtual CLBuildOptions generate_build_options(const ITensorInfo &input1, const ITensorInfo &input2, const ITensorInfo &output) = 0;
-
-    /** Generate the identifier for tuning
-     *
-     * @reutrn a string
-     */
-    virtual std::string generate_id_for_tuning(const std::string &kernel_name, const ITensorInfo &input1, const ITensorInfo &output) = 0;
-
-    /** Commmon configure function for element-wise operators with no additional options (e.g., Div, Min, Max, SquaredDiff)
-     *
-     */
-    void configure_common(ITensorInfo *input1, ITensorInfo *input2, ITensorInfo *output);
-    /** Commmon configure function for element-wise operators with no additional options (e.g., Div, Min, Max, SquaredDiff)
-     *
-     */
-    void configure_common(const CLCompileContext &compile_context, ITensorInfo *input1, ITensorInfo *input2, ITensorInfo *output);
-
-    ActivationLayerInfo _act_info;
-
-private:
-    const ITensorInfo *_input1; /**< Source tensor info 1 */
-    const ITensorInfo *_input2; /**< Source tensor info 2 */
-    ITensorInfo       *_output; /**< Destination tensor info */
-};
-
-/** Addition operation */
-class CLSaturatedArithmeticOperationKernel : public CLElementwiseOperationKernel
-{
-public:
-    CLSaturatedArithmeticOperationKernel()
-        : CLElementwiseOperationKernel(), _policy(), _op()
-    {
-    }
-
-    /** Static function to check if given info will lead to a valid configuration of @ref CLSaturatedArithmeticOperationKernel
-     *
-     * @param[in] op       Arithmetic operation to be executed.
-     * @param[in] input1   First tensor input info. Data types supported: U8/QASYMM8/QASYMM8_SIGNED/S16/QSYMM16/F16/S32/F32.
-     * @param[in] input2   Second tensor input info. Data types supported: Same as @p input1.
-     * @param[in] output   Output tensor info. Data types supported: Same as @p input1.
-     * @param[in] policy   Policy to use to handle overflow.
-     * @param[in] act_info (Optional) Activation layer information in case of a fused activation.
-     */
-    void configure(ArithmeticOperation op, ITensorInfo *input1, ITensorInfo *input2, ITensorInfo *output, const ConvertPolicy &policy, const ActivationLayerInfo &act_info = ActivationLayerInfo());
-    /** Static function to check if given info will lead to a valid configuration of @ref CLSaturatedArithmeticOperationKernel
-     *
-     * @param[in] compile_context The compile context to be used.
-     * @param[in] op              Arithmetic operation to be executed.
-     * @param[in] input1          First tensor input info. Data types supported: U8/QASYMM8/QASYMM8_SIGNED/S16/QSYMM16/F16/S32/F32.
-     * @param[in] input2          Second tensor input info. Data types supported: Same as @p input1.
-     * @param[in] output          Output tensor info. Data types supported: Same as @p input1.
-     * @param[in] policy          Policy to use to handle overflow.
-     * @param[in] act_info        (Optional) Activation layer information in case of a fused activation.
-     */
-    void configure(const CLCompileContext &compile_context, ArithmeticOperation op, ITensorInfo *input1, ITensorInfo *input2, ITensorInfo *output, const ConvertPolicy &policy,
-                   const ActivationLayerInfo &act_info = ActivationLayerInfo());
-
-    /** Static function to check if given info will lead to a valid configuration of @ref CLSaturatedArithmeticOperationKernel
-     *
-     * @param[in] op       Arithmetic operation to be executed.
-     * @param[in] input1   First tensor input info info. Data types supported: U8/QASYMM8/QASYMM8_SIGNED/S16/QSYMM16/F16/S32/F32.
-     * @param[in] input2   Second tensor input info info. Data types supported: Same as @p input1.
-     * @param[in] output   Output tensor info info. Data types supported: Same as @p input1.
-     * @param[in] policy   Policy to use to handle overflow.
-     * @param[in] act_info (Optional) Activation layer information in case of a fused activation.
-     *
-     * @return a Status
-     */
-    static Status validate(ArithmeticOperation op, const ITensorInfo *input1, const ITensorInfo *input2, const ITensorInfo *output, const ConvertPolicy &policy,
-                           const ActivationLayerInfo &act_info = ActivationLayerInfo());
-
-protected:
-    // Inherited methods overridden:
-    std::string name() override;
-    std::pair<Status, Window> validate_and_configure_window(ITensorInfo &input1, ITensorInfo &input2, ITensorInfo &output) override;
-    CLBuildOptions generate_build_options(const ITensorInfo &input1, const ITensorInfo &input2, const ITensorInfo &output) override;
-    std::string generate_id_for_tuning(const std::string &kernel_name, const ITensorInfo &input1, const ITensorInfo &output) override;
-
-private:
-    ConvertPolicy       _policy;
-    ArithmeticOperation _op;
-};
-
-class CLArithmeticOperationKernel : public CLElementwiseOperationKernel
-{
-public:
-    CLArithmeticOperationKernel()
-        : CLElementwiseOperationKernel(), _op()
-    {
-    }
-
-    /** Static function to check if given info will lead to a valid configuration of @ref CLArithmeticOperationKernel
-     *
-     * @param[in] op       Arithmetic operation to be executed.
-     * @param[in] input1   First tensor input info. Data types supported: U8/QASYMM8/QASYMM8_SIGNED/S16/QSYMM16/F16/S32/F32.
-     * @param[in] input2   Second tensor input info. Data types supported: Same as @p input1.
-     * @param[in] output   Output tensor info. Data types supported: Same as @p input1.
-     * @param[in] act_info (Optional) Activation layer information in case of a fused activation.
-     */
-    void configure(ArithmeticOperation op, ITensorInfo *input1, ITensorInfo *input2, ITensorInfo *output, const ActivationLayerInfo &act_info = ActivationLayerInfo());
-    /** Static function to check if given info will lead to a valid configuration of @ref CLArithmeticOperationKernel
-     *
-     * @param[in] compile_context The compile context to be used.
-     * @param[in] op              Arithmetic operation to be executed.
-     * @param[in] input1          First tensor input info. Data types supported: U8/QASYMM8/QASYMM8_SIGNED/S16/QSYMM16/F16/S32/F32.
-     * @param[in] input2          Second tensor input info. Data types supported: Same as @p input1.
-     * @param[in] output          Output tensor info. Data types supported: Same as @p input1.
-     * @param[in] act_info        (Optional) Activation layer information in case of a fused activation.
-     */
-    void configure(const CLCompileContext &compile_context, ArithmeticOperation op, ITensorInfo *input1, ITensorInfo *input2, ITensorInfo *output,
-                   const ActivationLayerInfo &act_info = ActivationLayerInfo());
-
-    /** Static function to check if given info will lead to a valid configuration of @ref CLArithmeticOperationKernel
-     *
-     * @param[in] op       Arithmetic operation to be executed.
-     * @param[in] input1   First tensor input info. Data types supported: U8/QASYMM8/QASYMM8_SIGNED/S16/QSYMM16/F16/S32/F32.
-     * @param[in] input2   Second tensor input info. Data types supported: Same as @p input1.
-     * @param[in] output   Output tensor info. Data types supported: Same as @p input1.
-     * @param[in] act_info (Optional) Activation layer information in case of a fused activation.
-     *
-     * @return a Status
-     */
-    static Status validate(ArithmeticOperation op, const ITensorInfo *input1, const ITensorInfo *input2, const ITensorInfo *output, const ActivationLayerInfo &act_info = ActivationLayerInfo());
-
-protected:
-    // Inherited methods overridden:
-    std::string name() override;
-    std::pair<Status, Window> validate_and_configure_window(ITensorInfo &input1, ITensorInfo &input2, ITensorInfo &output) override;
-    CLBuildOptions generate_build_options(const ITensorInfo &input1, const ITensorInfo &input2, const ITensorInfo &output) override;
-    std::string generate_id_for_tuning(const std::string &kernel_name, const ITensorInfo &input1, const ITensorInfo &output) override;
-
-private:
-    ArithmeticOperation _op;
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLELEMENTWISEOPERATIONKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLErodeKernel.h b/arm_compute/core/CL/kernels/CLErodeKernel.h
deleted file mode 100644
index 620201d625..0000000000
--- a/arm_compute/core/CL/kernels/CLErodeKernel.h
+++ /dev/null
@@ -1,59 +0,0 @@
-/*
- * Copyright (c) 2016-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLERODEKERNEL_H
-#define ARM_COMPUTE_CLERODEKERNEL_H
-
-#include "arm_compute/core/CL/ICLSimple2DKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the erode kernel.
- *
- */
-class CLErodeKernel : public ICLSimple2DKernel
-{
-public:
-    /**Initialise the kernel's input and output.
-     *
-     * @param[in]  input            An input tensor. Data types supported: U8
-     * @param[out] output           The output tensor. Data types supported: U8.
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, bool border_undefined);
-    /**Initialise the kernel's input and output.
-     *
-     * @param[in]  compile_context  The compile context to be used.
-     * @param[in]  input            An input tensor. Data types supported: U8
-     * @param[out] output           The output tensor. Data types supported: U8.
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, bool border_undefined);
-
-    // Inherited methods overridden:
-    BorderSize border_size() const override;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLERODEKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLFFTDigitReverseKernel.h b/arm_compute/core/CL/kernels/CLFFTDigitReverseKernel.h
deleted file mode 100644
index a196c8c64f..0000000000
--- a/arm_compute/core/CL/kernels/CLFFTDigitReverseKernel.h
+++ /dev/null
@@ -1,89 +0,0 @@
-/*
- * Copyright (c) 2019-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLFFTDIGITREVERSEKERNEL_H
-#define ARM_COMPUTE_CLFFTDIGITREVERSEKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-#include "arm_compute/core/KernelDescriptors.h"
-
-namespace arm_compute
-{
-// Forward declarations
-class ICLTensor;
-
-/** Interface for the digit reverse operation kernel. */
-class CLFFTDigitReverseKernel : public ICLKernel
-{
-public:
-    /** Constructor */
-    CLFFTDigitReverseKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLFFTDigitReverseKernel(const CLFFTDigitReverseKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLFFTDigitReverseKernel &operator=(const CLFFTDigitReverseKernel &) = delete;
-    /** Default Move Constructor. */
-    CLFFTDigitReverseKernel(CLFFTDigitReverseKernel &&) = default;
-    /** Default move assignment operator */
-    CLFFTDigitReverseKernel &operator=(CLFFTDigitReverseKernel &&) = default;
-    /** Default destructor */
-    ~CLFFTDigitReverseKernel() = default;
-    /** Set the input and output tensors.
-     *
-     * @param[in]  input  Source tensor. Data types supported: F32.
-     * @param[out] output Destination tensor. Data type supported: same as @p input
-     * @param[in]  idx    Digit reverse index tensor. Data type supported: U32
-     * @param[in]  config Kernel configuration.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, const ICLTensor *idx, const FFTDigitReverseKernelInfo &config);
-    /** Set the input and output tensors.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Source tensor. Data types supported: F32.
-     * @param[out] output          Destination tensor. Data type supported: same as @p input
-     * @param[in]  idx             Digit reverse index tensor. Data type supported: U32
-     * @param[in]  config          Kernel configuration.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, const ICLTensor *idx, const FFTDigitReverseKernelInfo &config);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLFFTDigitReverseKernel
-     *
-     * @param[in] input  Source tensor info. Data types supported: F32.
-     * @param[in] output Destination tensor info. Data type supported: same as @p input
-     * @param[in] idx    Digit reverse index tensor info. Data type supported: U32
-     * @param[in] config Kernel configuration.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output, const ITensorInfo *idx, const FFTDigitReverseKernelInfo &config);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;
-    ICLTensor       *_output;
-    const ICLTensor *_idx;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLFFTDIGITREVERSEKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLFFTRadixStageKernel.h b/arm_compute/core/CL/kernels/CLFFTRadixStageKernel.h
deleted file mode 100644
index d6d6067bc4..0000000000
--- a/arm_compute/core/CL/kernels/CLFFTRadixStageKernel.h
+++ /dev/null
@@ -1,97 +0,0 @@
-/*
- * Copyright (c) 2019-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLFFTRADIXSTAGEKERNEL_H
-#define ARM_COMPUTE_CLFFTRADIXSTAGEKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-#include "arm_compute/core/KernelDescriptors.h"
-
-#include <set>
-
-namespace arm_compute
-{
-// Forward declarations
-class ICLTensor;
-
-/** Interface for the FFT radix stage kernel. */
-class CLFFTRadixStageKernel : public ICLKernel
-{
-public:
-    /** Constructor */
-    CLFFTRadixStageKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLFFTRadixStageKernel(const CLFFTRadixStageKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLFFTRadixStageKernel &operator=(const CLFFTRadixStageKernel &) = delete;
-    /** Default Move Constructor. */
-    CLFFTRadixStageKernel(CLFFTRadixStageKernel &&) = default;
-    /** Default move assignment operator */
-    CLFFTRadixStageKernel &operator=(CLFFTRadixStageKernel &&) = default;
-    /** Default destructor */
-    ~CLFFTRadixStageKernel() = default;
-    /** Set the input and output tensors.
-     *
-     * @note If the output tensor is nullptr, the FFT will be performed in-place
-     *
-     * @param[in,out] input  Source tensor. Data types supported: F32.
-     * @param[out]    output Destination tensor. Can be nullptr. Data type supported: same as @p input
-     * @param[in]     config FFT descriptor metadata.
-     */
-    void configure(ICLTensor *input, ICLTensor *output, const FFTRadixStageKernelInfo &config);
-    /** Set the input and output tensors.
-     *
-     * @note If the output tensor is nullptr, the FFT will be performed in-place
-     *
-     * @param[in]     compile_context The compile context to be used.
-     * @param[in,out] input           Source tensor. Data types supported: F32.
-     * @param[out]    output          Destination tensor. Can be nullptr. Data type supported: same as @p input
-     * @param[in]     config          FFT descriptor metadata.
-     */
-    void configure(const CLCompileContext &compile_context, ICLTensor *input, ICLTensor *output, const FFTRadixStageKernelInfo &config);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLFFTRadixStageKernel
-     *
-     * @param[in] input  Source tensor info. Data types supported: F32.
-     * @param[in] output Destination tensor info. Can be nullptr. Data type supported: same as @p input
-     * @param[in] config FFT descriptor metadata.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output, const FFTRadixStageKernelInfo &config);
-    /** Returns the radix that are support by the FFT kernel
-     *
-     * @return A set of supported radix
-     */
-    static std::set<unsigned int> supported_radix();
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    ICLTensor *_input;
-    ICLTensor *_output;
-    bool       _run_in_place;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLFFTRADIXSTAGEKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLFFTScaleKernel.h b/arm_compute/core/CL/kernels/CLFFTScaleKernel.h
deleted file mode 100644
index c6dd176f58..0000000000
--- a/arm_compute/core/CL/kernels/CLFFTScaleKernel.h
+++ /dev/null
@@ -1,86 +0,0 @@
-/*
- * Copyright (c) 2019-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLFFTSCALEKERNEL_H
-#define ARM_COMPUTE_CLFFTSCALEKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-#include "arm_compute/core/KernelDescriptors.h"
-
-namespace arm_compute
-{
-// Forward declarations
-class ICLTensor;
-
-/** Interface for the inverse fft scale kernel. */
-class CLFFTScaleKernel : public ICLKernel
-{
-public:
-    /** Constructor */
-    CLFFTScaleKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLFFTScaleKernel(const CLFFTScaleKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLFFTScaleKernel &operator=(const CLFFTScaleKernel &) = delete;
-    /** Default Move Constructor. */
-    CLFFTScaleKernel(CLFFTScaleKernel &&) = default;
-    /** Default move assignment operator */
-    CLFFTScaleKernel &operator=(CLFFTScaleKernel &&) = default;
-    /** Default destructor */
-    ~CLFFTScaleKernel() = default;
-    /** Set the input and output tensors.
-     *
-     * @param[in,out] input  Source tensor. Data types supported: F32.
-     * @param[out]    output Destination tensor. Data type supported: same as @p input
-     * @param[in]     config Kernel configuration
-     */
-    void configure(ICLTensor *input, ICLTensor *output, const FFTScaleKernelInfo &config);
-    /** Set the input and output tensors.
-     *
-     * @param[in]     compile_context The compile context to be used.
-     * @param[in,out] input           Source tensor. Data types supported: F32.
-     * @param[out]    output          Destination tensor. Data type supported: same as @p input
-     * @param[in]     config          Kernel configuration
-     */
-    void configure(const CLCompileContext &compile_context, ICLTensor *input, ICLTensor *output, const FFTScaleKernelInfo &config);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLFFTScaleKernel
-     *
-     * @param[in] input  Source tensor info. Data types supported: F32.
-     * @param[in] output Destination tensor info. Data type supported: same as @p input
-     * @param[in] config Kernel configuration
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output, const FFTScaleKernelInfo &config);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    ICLTensor *_input;
-    ICLTensor *_output;
-    bool       _run_in_place;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLFFTSCALEKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLFastCornersKernel.h b/arm_compute/core/CL/kernels/CLFastCornersKernel.h
deleted file mode 100644
index 5d0da7d5d5..0000000000
--- a/arm_compute/core/CL/kernels/CLFastCornersKernel.h
+++ /dev/null
@@ -1,133 +0,0 @@
-/*
- * Copyright (c) 2016-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLFASTCORNERSKERNEL_H
-#define ARM_COMPUTE_CLFASTCORNERSKERNEL_H
-
-#include "arm_compute/core/CL/ICLArray.h"
-#include "arm_compute/core/CL/ICLKernel.h"
-#include "arm_compute/core/Types.h"
-
-#include <cstdint>
-
-namespace cl
-{
-class Buffer;
-}
-
-namespace arm_compute
-{
-class ICLTensor;
-using ICLImage = ICLTensor;
-
-/** CL kernel to perform fast corners */
-class CLFastCornersKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLFastCornersKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLFastCornersKernel(const CLFastCornersKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLFastCornersKernel &operator=(const CLFastCornersKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLFastCornersKernel(CLFastCornersKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLFastCornersKernel &operator=(CLFastCornersKernel &&) = default;
-    /** Default destructor */
-    ~CLFastCornersKernel() = default;
-
-    /** Initialise the kernel.
-     *
-     * @param[in]  input               Source image. Data types supported: U8.
-     * @param[out] output              Output image. Data types supported: U8.
-     * @param[in]  threshold           Threshold on difference between intensity of the central pixel and pixels on Bresenham's circle of radius 3.
-     * @param[in]  non_max_suppression True if non-maxima suppresion is applied, false otherwise.
-     * @param[in]  border_mode         Strategy to use for borders.
-     */
-    void configure(const ICLImage *input, ICLImage *output, float threshold, bool non_max_suppression, BorderMode border_mode);
-    /** Initialise the kernel.
-     *
-     * @param[in]  compile_context     The compile context to be used.
-     * @param[in]  input               Source image. Data types supported: U8.
-     * @param[out] output              Output image. Data types supported: U8.
-     * @param[in]  threshold           Threshold on difference between intensity of the central pixel and pixels on Bresenham's circle of radius 3.
-     * @param[in]  non_max_suppression True if non-maxima suppresion is applied, false otherwise.
-     * @param[in]  border_mode         Strategy to use for borders.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLImage *input, ICLImage *output, float threshold, bool non_max_suppression, BorderMode border_mode);
-
-    // Inherited methods overridden
-    void run(const Window &window, cl::CommandQueue &queue) override;
-    BorderSize border_size() const override;
-
-private:
-    const ICLImage *_input;
-    ICLImage       *_output;
-};
-
-/** CL kernel to copy keypoints information to ICLKeyPointArray and counts the number of key points */
-class CLCopyToArrayKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLCopyToArrayKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLCopyToArrayKernel(const CLCopyToArrayKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLCopyToArrayKernel &operator=(const CLCopyToArrayKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLCopyToArrayKernel(CLCopyToArrayKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLCopyToArrayKernel &operator=(CLCopyToArrayKernel &&) = default;
-    /** Default destructor */
-    ~CLCopyToArrayKernel() = default;
-
-    /** Initialise the kernel.
-     *
-     * @param[in]  input         Source image. Data types supported: U8.
-     * @param[in]  update_number Flag to indicate whether we need to update the number of corners
-     * @param[out] corners       Array of keypoints to store the results.
-     * @param[out] num_buffers   Number of keypoints to store the results.
-     */
-    void configure(const ICLImage *input, bool update_number, ICLKeyPointArray *corners, cl::Buffer *num_buffers);
-    /** Initialise the kernel.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Source image. Data types supported: U8.
-     * @param[in]  update_number   Flag to indicate whether we need to update the number of corners
-     * @param[out] corners         Array of keypoints to store the results.
-     * @param[out] num_buffers     Number of keypoints to store the results.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLImage *input, bool update_number, ICLKeyPointArray *corners, cl::Buffer *num_buffers);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLImage   *_input;      /**< source image */
-    ICLKeyPointArray *_corners;    /**< destination array */
-    cl::Buffer       *_num_buffer; /**< CL memory to record number of key points in the array */
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLFASTCORNERSKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLFillBorderKernel.h b/arm_compute/core/CL/kernels/CLFillBorderKernel.h
deleted file mode 100644
index 5323af4c0e..0000000000
--- a/arm_compute/core/CL/kernels/CLFillBorderKernel.h
+++ /dev/null
@@ -1,96 +0,0 @@
-/*
- * Copyright (c) 2016-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLFILLBORDERKERNEL_H
-#define ARM_COMPUTE_CLFILLBORDERKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-#include "arm_compute/core/PixelValue.h"
-#include "arm_compute/core/Types.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for filling the border of a kernel */
-class CLFillBorderKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLFillBorderKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLFillBorderKernel(const CLFillBorderKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLFillBorderKernel &operator=(const CLFillBorderKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLFillBorderKernel(CLFillBorderKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLFillBorderKernel &operator=(CLFillBorderKernel &&) = default;
-    /** Default destructor */
-    ~CLFillBorderKernel() = default;
-
-    /** Initialise the kernel's input, output and border mode.
-     *
-     * @param[in]     compile_context       The compile context to be used.
-     * @param[in,out] tensor                Tensor to process Data types supported: U8/QASYMM8/S8/QASYMM8_SIGNED/U16/S16/U32/S32/F16/F32.
-     * @param[in]     border_size           Size of the border to fill in elements.
-     * @param[in]     border_mode           Border mode to use for the convolution.
-     * @param[in]     constant_border_value (Optional) Constant value to use for borders if border_mode is set to CONSTANT.
-     */
-    void configure(const CLCompileContext &compile_context, ICLTensor *tensor, BorderSize border_size, BorderMode border_mode, const PixelValue &constant_border_value = PixelValue());
-    /** Initialise the kernel's input, output and border mode.
-     *
-     * @param[in,out] tensor                Tensor to process Data types supported: U8/QASYMM8/S8/QASYMM8_SIGNED/U16/S16/U32/S32/F16/F32.
-     * @param[in]     border_size           Size of the border to fill in elements.
-     * @param[in]     border_mode           Border mode to use for the convolution.
-     * @param[in]     constant_border_value (Optional) Constant value to use for borders if border_mode is set to CONSTANT.
-     */
-    void configure(ICLTensor *tensor, BorderSize border_size, BorderMode border_mode, const PixelValue &constant_border_value = PixelValue());
-    /** Initialise the kernel's input, output and border mode.
-     *
-     * @param[in]     compile_context       The compile context to be used.
-     * @param[in,out] tensor                Tensor to process Data types supported: U8/QASYMM8/S8/QASYMM8_SIGNED/U16/S16/U32/S32/F16/F32.
-     * @param[in]     border_size           Size of the border to fill in elements.
-     * @param[in]     border_mode           Border mode to use for the convolution.
-     * @param[in]     constant_border_value (Optional) Constant value to use for borders if border_mode is set to CONSTANT.
-     */
-    void configure(const CLCompileContext &compile_context, ITensorInfo *tensor, BorderSize border_size, BorderMode border_mode, const PixelValue &constant_border_value = PixelValue());
-
-    /** Function to set the constant value on fill border kernel depending on type.
-     *
-     * @param[in] idx                   Index of the kernel argument to set.
-     * @param[in] constant_border_value Constant value to use for borders if border_mode is set to CONSTANT.
-     */
-    template <class T>
-    void set_constant_border(unsigned int idx, const PixelValue &constant_border_value);
-
-    // Inherited methods overridden:
-    void run_op(ITensorPack &tensors, const Window &window, cl::CommandQueue &queue) override;
-    void run(const Window &window, cl::CommandQueue &queue) override;
-    bool is_parallelisable() const override;
-
-private:
-    ICLTensor *_tensor;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLFILLBORDERKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLFlattenLayerKernel.h b/arm_compute/core/CL/kernels/CLFlattenLayerKernel.h
deleted file mode 100644
index 4df0b33c8e..0000000000
--- a/arm_compute/core/CL/kernels/CLFlattenLayerKernel.h
+++ /dev/null
@@ -1,83 +0,0 @@
-/*
- * Copyright (c) 2018-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLFLATTENLAYERKERNEL_H
-#define ARM_COMPUTE_CLFLATTENLAYERKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** OpenCL interface for the flatten kernel.*/
-class CLFlattenLayerKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLFlattenLayerKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLFlattenLayerKernel(const CLFlattenLayerKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLFlattenLayerKernel &operator=(const CLFlattenLayerKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLFlattenLayerKernel(CLFlattenLayerKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLFlattenLayerKernel &operator=(CLFlattenLayerKernel &&) = default;
-    /** Set the input and output of the kernel.
-     *
-     * @param[in]  input  First input tensor to flatten with at least 3 dimensions.
-     *                    The dimensions above the third will be interpreted as batches. Data types supported: All.
-     * @param[out] output Output tensor with shape [w*h*d, input_batches] where:
-     *                    w = width input tensor, h = height input tensor and d = depth input tensor. Data type supported: same as @p input
-     */
-    void configure(const ICLTensor *input, ICLTensor *output);
-    /** Set the input and output of the kernel.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           First input tensor to flatten with at least 3 dimensions.
-     *                             The dimensions above the third will be interpreted as batches. Data types supported: All.
-     * @param[out] output          Output tensor with shape [w*h*d, input_batches] where:
-     *                    w = width input tensor, h = height input tensor and d = depth input tensor. Data type supported: same as @p input
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLFlattenLayerKernel
-     *
-     * @param[in]  input  First input tensor to flatten with at least 3 dimensions.
-     *                    The dimensions above the third will be interpreted as batches. Data types supported: All.
-     * @param[out] output Output tensor with shape [w*h*d, input_batches] where:
-     *                    w = width input tensor, h = height input tensor and d = depth input tensor. Data type supported: same as @p input
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-public:
-    const ICLTensor *_input;
-    ICLTensor       *_output;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLFLATTENLAYERKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLFloorKernel.h b/arm_compute/core/CL/kernels/CLFloorKernel.h
deleted file mode 100644
index 3b1d3f10cc..0000000000
--- a/arm_compute/core/CL/kernels/CLFloorKernel.h
+++ /dev/null
@@ -1,81 +0,0 @@
-/*
- * Copyright (c) 2017-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLFLOORKERNEL_H
-#define ARM_COMPUTE_CLFLOORKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** OpenCL kernel to perform a floor operation */
-class CLFloorKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLFloorKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLFloorKernel(const CLFloorKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLFloorKernel &operator=(const CLFloorKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLFloorKernel(CLFloorKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLFloorKernel &operator=(CLFloorKernel &&) = default;
-    /** Default destructor */
-    ~CLFloorKernel() = default;
-    /** Set the source, destination of the kernel
-     *
-     * @param[in]  input  Source tensor. Data type supported: F16/F32.
-     * @param[out] output Destination tensor. Same as @p input
-     */
-    void configure(const ICLTensor *input, ICLTensor *output);
-
-    /** Set the source, destination of the kernel
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Source tensor. Data type supported: F16/F32.
-     * @param[out] output          Destination tensor. Same as @p input
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output);
-
-    /** Static function to check if given info will lead to a valid configuration of @ref CLFloorKernel
-     *
-     * @param[in] input  Source tensor info. Data type supported: F16/F32.
-     * @param[in] output Destination tensor info. Same as @p input
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;
-    ICLTensor       *_output;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLFLOORKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLFuseBatchNormalizationKernel.h b/arm_compute/core/CL/kernels/CLFuseBatchNormalizationKernel.h
deleted file mode 100644
index 3ec251c858..0000000000
--- a/arm_compute/core/CL/kernels/CLFuseBatchNormalizationKernel.h
+++ /dev/null
@@ -1,126 +0,0 @@
-/*
- * Copyright (c) 2018-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLFUSEBATCHNORMALIZATIONKERNEL_H
-#define ARM_COMPUTE_CLFUSEBATCHNORMALIZATIONKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-// Forward declarations
-class ICLTensor;
-
-/** OpenCL kernel to fuse the batch normalization node to a preceding convolution node */
-class CLFuseBatchNormalizationKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLFuseBatchNormalizationKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLFuseBatchNormalizationKernel(const CLFuseBatchNormalizationKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLFuseBatchNormalizationKernel &operator=(const CLFuseBatchNormalizationKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLFuseBatchNormalizationKernel(CLFuseBatchNormalizationKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLFuseBatchNormalizationKernel &operator=(CLFuseBatchNormalizationKernel &&) = default;
-    /** Default destructor */
-    ~CLFuseBatchNormalizationKernel() = default;
-    /** Set the source, destination of the kernel
-     *
-     * @param[in]  input_weights Input weights tensor for convolution or depthwise convolution layer. Data type supported: F16/F32. Data layout supported: NCHW, NHWC
-     * @param[in]  bn_mean       Batch normalization layer mean tensor. Same as @p input_weights
-     * @param[in]  bn_var        Batch normalization layer variance tensor. Same as @p input_weights
-     * @param[out] fused_weights Output fused weights tensor. It can be a nullptr in case of in-place computation. Same as @p input_weights
-     * @param[out] fused_bias    Output fused bias tensor. It can be a nullptr in case of in-place computation and input_bias != nullptr. Same as @p input_weights
-     * @param[in]  input_bias    (Optional) Input bias tensor for convolution or depthwise convolution layer. It can be a nullptr in case the bias tensor is not required. Same as @p input_weights
-     * @param[in]  bn_beta       (Optional) Batch normalization layer beta tensor. It can be a nullptr in case the beta tensor is not required. Same as @p input_weights
-     *                           @note if nullptr, bn_beta is set to 0.0
-     * @param[in]  bn_gamma      (Optional) Batch normalization layer gamma tensor. It can be a nullptr in case the gamma tensor is not required. Same as @p input_weights
-     *                           @note if nullptr, bn_gamma is set to 1.0
-     * @param[in]  epsilon       (Optional) Batch normalization layer epsilon parameter. Defaults to 0.001f.
-     * @param[in]  fbn_type      (Optional) Fused batch normalization type. Defaults to CONVOLUTION.
-     */
-    void configure(const ICLTensor *input_weights, const ICLTensor *bn_mean, const ICLTensor *bn_var, ICLTensor *fused_weights, ICLTensor *fused_bias,
-                   const ICLTensor *input_bias = nullptr, const ICLTensor *bn_beta = nullptr, const ICLTensor *bn_gamma = nullptr,
-                   float epsilon = 0.001f, FuseBatchNormalizationType fbn_type = FuseBatchNormalizationType::CONVOLUTION);
-    /** Set the source, destination of the kernel
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input_weights   Input weights tensor for convolution or depthwise convolution layer. Data type supported: F16/F32. Data layout supported: NCHW, NHWC
-     * @param[in]  bn_mean         Batch normalization layer mean tensor. Same as @p input_weights
-     * @param[in]  bn_var          Batch normalization layer variance tensor. Same as @p input_weights
-     * @param[out] fused_weights   Output fused weights tensor. It can be a nullptr in case of in-place computation. Same as @p input_weights
-     * @param[out] fused_bias      Output fused bias tensor. It can be a nullptr in case of in-place computation and input_bias != nullptr. Same as @p input_weights
-     * @param[in]  input_bias      (Optional) Input bias tensor for convolution or depthwise convolution layer. It can be a nullptr in case the bias tensor is not required. Same as @p input_weights
-     * @param[in]  bn_beta         (Optional) Batch normalization layer beta tensor. It can be a nullptr in case the beta tensor is not required. Same as @p input_weights
-     *                             @note if nullptr, bn_beta is set to 0.0
-     * @param[in]  bn_gamma        (Optional) Batch normalization layer gamma tensor. It can be a nullptr in case the gamma tensor is not required. Same as @p input_weights
-     *                             @note if nullptr, bn_gamma is set to 1.0
-     * @param[in]  epsilon         (Optional) Batch normalization layer epsilon parameter. Defaults to 0.001f.
-     * @param[in]  fbn_type        (Optional) Fused batch normalization type. Defaults to CONVOLUTION.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input_weights, const ICLTensor *bn_mean, const ICLTensor *bn_var, ICLTensor *fused_weights, ICLTensor *fused_bias,
-                   const ICLTensor *input_bias = nullptr, const ICLTensor *bn_beta = nullptr, const ICLTensor *bn_gamma = nullptr,
-                   float epsilon = 0.001f, FuseBatchNormalizationType fbn_type = FuseBatchNormalizationType::CONVOLUTION);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLFuseBatchNormalizationKernel
-     *
-     * @param[in] input_weights Input weights tensor info for convolution or depthwise convolution layer. Data type supported: F16/F32. Data layout supported: NCHW, NHWC
-     * @param[in] bn_mean       Batch normalization layer mean tensor info. Same as @p input_weights
-     * @param[in] bn_var        Batch normalization layer variance tensor info. Same as @p input_weights
-     * @param[in] fused_weights Output fused weights tensor info. It can be a nullptr in case of in-place computation. Same as @p input_weights
-     * @param[in] fused_bias    Output fused bias tensor info. It can be a nullptr in case of in-place computation and input_bias != nullptr. Same as @p input_weights
-     * @param[in] input_bias    (Optional) Input bias tensor info for convolution or depthwise convolution layer. It can be a nullptr in case the bias tensor is not required. Same as @p input_weights
-     * @param[in] bn_beta       (Optional) Batch normalization layer beta tensor info. It can be a nullptr in case the beta tensor is not required. Same as @p input_weights
-     *                          @note if nullptr, bn_beta is set to 0.0
-     * @param[in] bn_gamma      (Optional) Batch normalization layer gamma tensor info. It can be a nullptr in case the gamma tensor is not required. Same as @p input_weights
-     *                          @note if nullptr, bn_gamma is set to 1.0
-     * @param[in] epsilon       (Optional) Batch normalization layer epsilon parameter. Defaults to 0.001f.
-     * @param[in] fbn_type      (Optional) Fused batch normalization type. Defaults to CONVOLUTION.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input_weights, const ITensorInfo *bn_mean, const ITensorInfo *bn_var,
-                           const ITensorInfo *fused_weights, const ITensorInfo *fused_bias,
-                           const ITensorInfo *input_bias = nullptr, const ITensorInfo *bn_beta = nullptr, const ITensorInfo *bn_gamma = nullptr,
-                           float epsilon = 0.001f, FuseBatchNormalizationType fbn_type = FuseBatchNormalizationType::CONVOLUTION);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input_weights;
-    const ICLTensor *_input_bias;
-    const ICLTensor *_bn_mean;
-    const ICLTensor *_bn_var;
-    const ICLTensor *_bn_gamma;
-    const ICLTensor *_bn_beta;
-    ICLTensor       *_fused_weights;
-    ICLTensor       *_fused_bias;
-    float            _epsilon;
-    bool             _run_in_place_weights;
-    bool             _run_in_place_bias;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLFUSEBATCHNORMALIZATIONKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLGEMMLowpMatrixMultiplyNativeKernel.h b/arm_compute/core/CL/kernels/CLGEMMLowpMatrixMultiplyNativeKernel.h
deleted file mode 100644
index 2d5e4a3346..0000000000
--- a/arm_compute/core/CL/kernels/CLGEMMLowpMatrixMultiplyNativeKernel.h
+++ /dev/null
@@ -1,108 +0,0 @@
-/*
- * Copyright (c) 2019-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLGEMMLOWPMATRIXMULTIPLYNATIVEKERNEL_H
-#define ARM_COMPUTE_CLGEMMLOWPMATRIXMULTIPLYNATIVEKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** OpenCL kernel to multiply matrices with QASYMM8/QASYMM8_SIGNED data type */
-class CLGEMMLowpMatrixMultiplyNativeKernel : public ICLKernel
-{
-public:
-    /** Default Constructor */
-    CLGEMMLowpMatrixMultiplyNativeKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLGEMMLowpMatrixMultiplyNativeKernel(const CLGEMMLowpMatrixMultiplyNativeKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLGEMMLowpMatrixMultiplyNativeKernel &operator=(const CLGEMMLowpMatrixMultiplyNativeKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLGEMMLowpMatrixMultiplyNativeKernel(CLGEMMLowpMatrixMultiplyNativeKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLGEMMLowpMatrixMultiplyNativeKernel &operator=(CLGEMMLowpMatrixMultiplyNativeKernel &&) = default;
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in]  input0    Input tensor containing the LHS matrix. Data type supported: QASYMM8/QASYMM8_SIGNED
-     * @param[in]  input1    Input tensor containing the RHS matrix. Data type supported: same as @p input0
-     * @param[out] output    Output tensor to store the result of matrix multiplication. Data type supported: S32
-     * @param[in]  lhs_info  LHS matrix information used to retrieve the number of rows to be processed by each thread
-     *                       lhs_info.m0: 2,3,4,5,6,7,8
-     *                       lhs_info.k0: 2,3,4,8,16
-     * @param[in]  rhs_info  RHS matrix information used to retrieve the number of columns to be processed by each thread
-     *                       rhs_info.n0: 2,3,4,8,16
-     *                       rhs_info.k0: same as lhs_info.k0
-     * @param[in]  gemm_info GEMM information used to retrieve the original dimensions of the input matrices
-     */
-    void configure(const ICLTensor *input0, const ICLTensor *input1, ICLTensor *output, const GEMMLHSMatrixInfo &lhs_info, const GEMMRHSMatrixInfo &rhs_info, const GEMMReshapeInfo &gemm_info);
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input0          Input tensor containing the LHS matrix. Data type supported: QASYMM8/QASYMM8_SIGNED
-     * @param[in]  input1          Input tensor containing the RHS matrix. Data type supported: same as @p input0
-     * @param[out] output          Output tensor to store the result of matrix multiplication. Data type supported: S32
-     * @param[in]  lhs_info        LHS matrix information used to retrieve the number of rows to be processed by each thread
-     *                             lhs_info.m0: 2,3,4,5,6,7,8
-     *                             lhs_info.k0: 2,3,4,8,16
-     * @param[in]  rhs_info        RHS matrix information used to retrieve the number of columns to be processed by each thread
-     *                             rhs_info.n0: 2,3,4,8,16
-     *                             rhs_info.k0: same as lhs_info.k0
-     * @param[in]  gemm_info       GEMM information used to retrieve the original dimensions of the input matrices
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input0, const ICLTensor *input1, ICLTensor *output, const GEMMLHSMatrixInfo &lhs_info, const GEMMRHSMatrixInfo &rhs_info,
-                   const GEMMReshapeInfo &gemm_info);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLGEMMLowpMatrixMultiplyNativeKernel
-     *
-     * @param[in] input0    Input tensor info for the LHS matrix. Data type supported: QASYMM8/QASYMM8_SIGNED
-     * @param[in] input1    Input tensor info for the RHS matrix. Data type supported: same as @p input0
-     * @param[in] output    Output tensor info. Data type supported: S32
-     * @param[in] lhs_info  LHS matrix information used to retrieve the number of rows to be processed by each thread
-     *                      lhs_info.m0: 2,3,4,5,6,7,8
-     *                      lhs_info.k0: 2,3,4,8,16
-     * @param[in] rhs_info  RHS matrix information used to retrieve the number of columns to be processed by each thread
-     *                      rhs_info.n0: 2,3,4,8,16
-     *                      rhs_info.k0: same as lhs_info.k0
-     * @param[in] gemm_info GEMM information used to retrieve the original dimensions of the input matrices
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input0, const ITensorInfo *input1, const ITensorInfo *output, const GEMMLHSMatrixInfo &lhs_info, const GEMMRHSMatrixInfo &rhs_info,
-                           const GEMMReshapeInfo &gemm_info);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input0;
-    const ICLTensor *_input1;
-    ICLTensor       *_output;
-    bool             _slide_matrix_b;
-    bool             _reinterpret_input_as_3d;
-    bool             _reinterpret_output_as_3d;
-    bool             _use_dummy_work_items;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLGEMMLOWPMATRIXMULTIPLYNATIVEKERNEL_H*/
diff --git a/arm_compute/core/CL/kernels/CLGEMMLowpMatrixMultiplyReshapedKernel.h b/arm_compute/core/CL/kernels/CLGEMMLowpMatrixMultiplyReshapedKernel.h
deleted file mode 100644
index f2eb447834..0000000000
--- a/arm_compute/core/CL/kernels/CLGEMMLowpMatrixMultiplyReshapedKernel.h
+++ /dev/null
@@ -1,123 +0,0 @@
-/*
- * Copyright (c) 2019-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLGEMMLOWPMATRIXMULTIPLYRESHAPEDKERNEL_H
-#define ARM_COMPUTE_CLGEMMLOWPMATRIXMULTIPLYRESHAPEDKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** OpenCL kernel to multiply matrices when both the input matrices LHS (input0) and RHS (input1) have been reshaped
- *
- * @note The input matrices @p input0 and @p input1 must be reshaped through @ref CLGEMMReshapeLHSMatrixKernel and  @ref CLGEMMReshapeRHSMatrixKernel
- */
-class CLGEMMLowpMatrixMultiplyReshapedKernel : public ICLKernel
-{
-public:
-    /** Default Constructor */
-    CLGEMMLowpMatrixMultiplyReshapedKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLGEMMLowpMatrixMultiplyReshapedKernel(const CLGEMMLowpMatrixMultiplyReshapedKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLGEMMLowpMatrixMultiplyReshapedKernel &operator=(const CLGEMMLowpMatrixMultiplyReshapedKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLGEMMLowpMatrixMultiplyReshapedKernel(CLGEMMLowpMatrixMultiplyReshapedKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLGEMMLowpMatrixMultiplyReshapedKernel &operator=(CLGEMMLowpMatrixMultiplyReshapedKernel &&) = default;
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in]  input0    Input tensor containing the LHS reshaped matrix. Data type supported: QASYMM8/QASYMM8_SIGNED. The number of dimensions for the LHS matrix must be less or equal than 4.
-     * @param[in]  input1    Input tensor containing the RHS reshaped matrix. Data type supported: same as @p input0. The number of dimensions for the RHS matrix must be less or equal than 3.
-     * @param[out] output    Output tensor to store the result of matrix multiplication. Data type supported: S32
-     * @param[in]  lhs_info  LHS matrix information used for reshaping the input0 tensor.  Only the following values are supported:
-     *                       lhs_info.m0: 2,3,4,5,6,7,8
-     *                       lhs_info.k0: 2,3,4,8,16
-     *                       lhs_info.transpose: false
-     * @param[in]  rhs_info  RHS matrix information used for reshaping the input1 tensor.  Only the following values are supported:
-     *                       rhs_info.n0: 2,3,4,8,16
-     *                       rhs_info.k0: same as lhs_info.k0
-     *                       rhs_info.transpose: true
-     * @param[in]  gemm_info GEMM information used to retrieve the original dimensions of the input matrices
-     *
-     * @note lhs_info.k0 must be equal to rhs_info.k0
-     */
-    void configure(const ICLTensor *input0, const ICLTensor *input1, ICLTensor *output, const GEMMLHSMatrixInfo &lhs_info, const GEMMRHSMatrixInfo &rhs_info, const GEMMReshapeInfo &gemm_info);
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input0          Input tensor containing the LHS reshaped matrix. Data type supported: QASYMM8/QASYMM8_SIGNED. The number of dimensions for the LHS matrix must be less or equal than 4.
-     * @param[in]  input1          Input tensor containing the RHS reshaped matrix. Data type supported: same as @p input0. The number of dimensions for the RHS matrix must be less or equal than 3.
-     * @param[out] output          Output tensor to store the result of matrix multiplication. Data type supported: S32
-     * @param[in]  lhs_info        LHS matrix information used for reshaping the input0 tensor.  Only the following values are supported:
-     *                             lhs_info.m0: 2,3,4,5,6,7,8
-     *                             lhs_info.k0: 2,3,4,8,16
-     *                             lhs_info.transpose: false
-     * @param[in]  rhs_info        RHS matrix information used for reshaping the input1 tensor.  Only the following values are supported:
-     *                             rhs_info.n0: 2,3,4,8,16
-     *                             rhs_info.k0: same as lhs_info.k0
-     *                             rhs_info.transpose: true
-     * @param[in]  gemm_info       GEMM information used to retrieve the original dimensions of the input matrices
-     *
-     * @note lhs_info.k0 must be equal to rhs_info.k0
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input0, const ICLTensor *input1, ICLTensor *output, const GEMMLHSMatrixInfo &lhs_info, const GEMMRHSMatrixInfo &rhs_info,
-                   const GEMMReshapeInfo &gemm_info);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLGEMMLowpMatrixMultiplyReshapedKernel
-     *
-     * @param[in] input0    Input tensor info containing the LHS reshaped matrix. Data type supported: QASYMM8/QASYMM8_SIGNED. The number of dimensions for the LHS matrix must be less or equal than 4.
-     * @param[in] input1    Input tensor info containing the RHS reshaped matrix. Data type supported: same as @p input0. The number of dimensions for the RHS matrix must be less or equal than 3.
-     * @param[in] output    Output tensor info. Data type supported: S32
-     * @param[in] lhs_info  LHS matrix information used for reshaping the input0 tensor.  Only the following values are supported:
-     *                      lhs_info.m0: 2,3,4,5,6,7,8
-     *                      lhs_info.k0: 2,3,4,8,16
-     *                      lhs_info.transpose: false
-     * @param[in] rhs_info  RHS matrix information used for reshaping the input1 tensor.  Only the following values are supported:
-     *                      rhs_info.n0: 2,3,4,8,16
-     *                      rhs_info.k0: 2,3,4,8,16
-     *                      rhs_info.transpose: true
-     * @param[in] gemm_info GEMM information used to retrieve the original dimensions of the input matrices
-     *
-     * @note lhs_info.k0 must be equal to rhs_info.k0
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input0, const ITensorInfo *input1, const ITensorInfo *output, const GEMMLHSMatrixInfo &lhs_info, const GEMMRHSMatrixInfo &rhs_info,
-                           const GEMMReshapeInfo &gemm_info);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input0;
-    const ICLTensor *_input1;
-    ICLTensor       *_output;
-    bool             _slide_matrix_b;
-    bool             _reinterpret_output_as_3d;
-    unsigned int     _k;
-    bool             _use_dummy_work_items;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLGEMMLOWPMATRIXMULTIPLYRESHAPEDKERNEL_H*/
diff --git a/arm_compute/core/CL/kernels/CLGEMMLowpMatrixMultiplyReshapedOnlyRHSKernel.h b/arm_compute/core/CL/kernels/CLGEMMLowpMatrixMultiplyReshapedOnlyRHSKernel.h
deleted file mode 100644
index a2295143de..0000000000
--- a/arm_compute/core/CL/kernels/CLGEMMLowpMatrixMultiplyReshapedOnlyRHSKernel.h
+++ /dev/null
@@ -1,155 +0,0 @@
-/*
- * Copyright (c) 2019-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLGEMMLOWPMATRIXMULTIPLYRESHAPEDONLYRHSKERNEL_H
-#define ARM_COMPUTE_CLGEMMLOWPMATRIXMULTIPLYRESHAPEDONLYRHSKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-#include "arm_compute/core/KernelDescriptors.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** OpenCL kernel to multiply matrices with QASYMM8 data type when only the input matrix RHS (input1) has been reshaped
- *
- * @note The input matrix input1 must be reshaped through @ref CLGEMMReshapeRHSMatrixKernel
- * @note For fused output stage, only GEMMLowpOutputStageType::QUANTIZE_DOWN_FIXEDPOINT type is supported
- */
-class CLGEMMLowpMatrixMultiplyReshapedOnlyRHSKernel : public ICLKernel
-{
-public:
-    /** Default Constructor */
-    CLGEMMLowpMatrixMultiplyReshapedOnlyRHSKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLGEMMLowpMatrixMultiplyReshapedOnlyRHSKernel(const CLGEMMLowpMatrixMultiplyReshapedOnlyRHSKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLGEMMLowpMatrixMultiplyReshapedOnlyRHSKernel &operator=(const CLGEMMLowpMatrixMultiplyReshapedOnlyRHSKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLGEMMLowpMatrixMultiplyReshapedOnlyRHSKernel(CLGEMMLowpMatrixMultiplyReshapedOnlyRHSKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLGEMMLowpMatrixMultiplyReshapedOnlyRHSKernel &operator=(CLGEMMLowpMatrixMultiplyReshapedOnlyRHSKernel &&) = default;
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in]  input0             Input tensor containing the LHS matrix. Data type supported: QASYMM8/QASYMM8_SIGNED
-     * @param[in]  input1             Input tensor containing the RHS reshaped matrix. Data type supported: QASYMM8/QASYMM8_SIGNED/QSYMM8/QSYMM8_PER_CHANNEL
-     * @param[out] output             Output tensor. Data type supported: QASYMM8/QASYMM8_SIGNED/S32.
-     * @param[in]  gemm_info          GEMM information used to retrieve the original dimensions of the input matrices, output stage information and RHS/LHS info.
-     *                                Only the following values are supported for LHS info:
-     *                                lhs_info.m0: 2,3,4,5,6,7,8
-     *                                lhs_info.k0: 2,3,4,8,16
-     *                                Only the following values are supported for RHS info:
-     *                                rhs_info.n0: 2,3,4,8,16
-     *                                rhs_info.k0: same as lhs_info.k0
-     *                                rhs_info.transpose: true
-     * @param[in]  vector_sum_col     (Optional) Input row-vector of sums of all the entries in each column of matrix B.
-     *                                Note: vector_sum_col can be a nullptr in case a_offset = 0. Data type supported: S32
-     * @param[in]  vector_sum_row     (Optional) Input row-vector of sums of all the entries in each row of matrix A.
-     *                                Note: vector_sum_row can be a nullptr in case b_offset = 0. Data type supported: S32
-     * @param[in]  bias               (Optional) Biases tensor. Only shared biases supported and it can be a nullptr if the addition of biases is not required.
-     *                                Biases are 1D tensor with dimensions [OFM]. Data type supported: S32.
-     * @param[in]  output_multipliers (Optional) Output multipliers tensor. In case of per-channel quantization, the number of multipliers must be equal to the number of filters (OFM).
-     *                                Supported data types: S32.
-     * @param[in]  output_shifts      (Optional) Output shifts tensor. In case of per-channel quantization, the number of multipliers must be equal to the number of filters (OFM).
-     *                                Supported data types: S32.
-     */
-    void configure(const ICLTensor *input0, const ICLTensor *input1, ICLTensor *output, const GEMMKernelInfo &gemm_info, const ICLTensor *vector_sum_col = nullptr,
-                   const ICLTensor *vector_sum_row = nullptr, const ICLTensor *bias = nullptr, const ICLTensor *output_multipliers = nullptr, const ICLTensor *output_shifts = nullptr);
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in]  compile_context    The compile context to be used.
-     * @param[in]  input0             Input tensor containing the LHS matrix. Data type supported: QASYMM8/QASYMM8_SIGNED
-     * @param[in]  input1             Input tensor containing the RHS reshaped matrix. Data type supported: same as @p input0
-     * @param[out] output             Output tensor. Data type supported: QASYMM8/QASYMM8_SIGNED/S32.
-     * @param[in]  gemm_info          GEMM information used to retrieve the original dimensions of the input matrices, output stage information and RHS/LHS info.
-     *                                Only the following values are supported for LHS info:
-     *                                lhs_info.m0: 2,3,4,5,6,7,8
-     *                                lhs_info.k0: 2,3,4,8,16
-     *                                Only the following values are supported for RHS info:
-     *                                rhs_info.n0: 2,3,4,8,16
-     *                                rhs_info.k0: same as lhs_info.k0
-     *                                rhs_info.transpose: true
-     * @param[in]  vector_sum_col     (Optional) Input row-vector of sums of all the entries in each column of matrix B.
-     *                                Note: vector_sum_col can be a nullptr in case a_offset = 0. Data type supported: S32
-     * @param[in]  vector_sum_row     (Optional) Input row-vector of sums of all the entries in each row of matrix A.
-     *                                Note: vector_sum_row can be a nullptr in case b_offset = 0. Data type supported: S32
-     * @param[in]  bias               (Optional) Biases tensor. Only shared biases supported and it can be a nullptr if the addition of biases is not required.
-     *                                Biases are 1D tensor with dimensions [OFM]. Data type supported: S32.
-     * @param[in]  output_multipliers (Optional) Output multipliers tensor. In case of per-channel quantization, the number of multipliers must be equal to the number of filters (OFM).
-     *                                Supported data types: S32.
-     * @param[in]  output_shifts      (Optional) Output shifts tensor. In case of per-channel quantization, the number of multipliers must be equal to the number of filters (OFM).
-     *                                Supported data types: S32.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input0, const ICLTensor *input1, ICLTensor *output, const GEMMKernelInfo &gemm_info, const ICLTensor *vector_sum_col = nullptr,
-                   const ICLTensor *vector_sum_row = nullptr, const ICLTensor *bias = nullptr, const ICLTensor *output_multipliers = nullptr, const ICLTensor *output_shifts = nullptr);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLGEMMLowpMatrixMultiplyReshapedOnlyRHSKernel
-     *
-     * @param[in] input0             Input tensor info for the LHS matrix. Data type supported: QASYMM8/QASYMM8_SIGNED
-     * @param[in] input1             Input tensor info for the RHS reshaped matrix. Data type supported: QASYMM8/QASYMM8_SIGNED/QSYMM8/QSYMM8_PER_CHANNEL
-     * @param[in] output             Output tensor info. Data type supported: QASYMM8/QASYMM8_SIGNED/S32.
-     * @param[in] gemm_info          GEMM information used to retrieve the original dimensions of the input matrices, output stage information and RHS/LHS info.
-     *                               Only the following values are supported for LHS info:
-     *                               lhs_info.m0: 2,3,4,5,6,7,8
-     *                               lhs_info.k0: 2,3,4,8,16
-     *                               Only the following values are supported for RHS info:
-     *                               rhs_info.n0: 2,3,4,8,16
-     *                               rhs_info.k0: same as lhs_info.k0
-     *                               rhs_info.transpose: true
-     * @param[in] vector_sum_col     (Optional) Input row-vector info of sums of all the entries in each column of matrix B.
-     *                               Note: vector_sum_col can be a nullptr in case a_offset = 0. Data type supported: S32
-     * @param[in] vector_sum_row     (Optional) Input row-vector info of sums of all the entries in each row of matrix A.
-     *                               Note: vector_sum_row can be a nullptr in case b_offset = 0. Data type supported: S32
-     * @param[in] bias               (Optional) Biases tensor info. Only shared biases supported and it can be a nullptr if the addition of biases is not required.
-     *                               Biases are 1D tensor with dimensions [OFM]. Data type supported: S32.
-     * @param[in] output_multipliers (Optional) Output multipliers tensor info. In case of per-channel quantization, the number of multipliers must be equal to the number of filters (OFM).
-     *                               Supported data types: S32.
-     * @param[in] output_shifts      (Optional) Output shifts tensor info. In case of per-channel quantization, the number of multipliers must be equal to the number of filters (OFM).
-     *                               Supported data types: S32.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input0, const ITensorInfo *input1, const ITensorInfo *output, const GEMMKernelInfo &gemm_info, const ITensorInfo *vector_sum_col = nullptr,
-                           const ITensorInfo *vector_sum_row = nullptr, const ITensorInfo *bias = nullptr, const ITensorInfo *output_multipliers = nullptr,
-                           const ITensorInfo *output_shifts = nullptr);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input0;
-    const ICLTensor *_input1;
-    ICLTensor       *_output;
-    const ICLTensor *_vector_sum_col;
-    const ICLTensor *_vector_sum_row;
-    const ICLTensor *_bias;
-    const ICLTensor *_output_multipliers;
-    const ICLTensor *_output_shifts;
-    bool             _slide_matrix_b;
-    bool             _reinterpret_input_as_3d;
-    bool             _reinterpret_output_as_3d;
-    bool             _use_dummy_work_items;
-    bool             _is_quantized_per_channel;
-    bool             _fuse_output_stage;
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLGEMMLOWPMATRIXMULTIPLYRESHAPEDONLYRHSKERNEL_H */
\ No newline at end of file
diff --git a/arm_compute/core/CL/kernels/CLGEMMLowpOffsetContributionKernel.h b/arm_compute/core/CL/kernels/CLGEMMLowpOffsetContributionKernel.h
deleted file mode 100644
index 1d3b3110b3..0000000000
--- a/arm_compute/core/CL/kernels/CLGEMMLowpOffsetContributionKernel.h
+++ /dev/null
@@ -1,116 +0,0 @@
-/*
- * Copyright (c) 2017-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLGEMMLOWPOFFSETCONTRIBUTIONKERNEL_H
-#define ARM_COMPUTE_CLGEMMLOWPOFFSETCONTRIBUTIONKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** OpenCL kernel used to add the offset contribution after the matrix multiplication. The computation is performed in-place
- *
- * This kernel takes a final int32 accumulator value (the output of the matrix multiplication),
- * and adds to it the offset contribution of matrix A and matrix B in-place.
- *
- * The final result is:
- *
- * mm_result[i][k] = mm_result[i][k] +
- *                   (vector_sum_col[k] * a_offset) +
- *                   (vector_sum_row[i] * b_offset) +
- *                   (a_offset * b_offset * k)
- *
- */
-class CLGEMMLowpOffsetContributionKernel : public ICLKernel
-{
-public:
-    /** Constructor */
-    CLGEMMLowpOffsetContributionKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers)*/
-    CLGEMMLowpOffsetContributionKernel(const CLGEMMLowpOffsetContributionKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers)*/
-    CLGEMMLowpOffsetContributionKernel &operator=(const CLGEMMLowpOffsetContributionKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLGEMMLowpOffsetContributionKernel(CLGEMMLowpOffsetContributionKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLGEMMLowpOffsetContributionKernel &operator=(CLGEMMLowpOffsetContributionKernel &&) = default;
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in, out] mm_result      Input tensor containing the result of the matrix multiplication. Data type supported: S32
-     * @param[in]      vector_sum_col Input row-vector of sums of all the entries in each column of matrix B.
-     *                                Note: vector_sum_col can be a nullptr in case a_offset = 0. Data type supported: same as @p mm_result
-     * @param[in]      vector_sum_row Input row-vector of sums of all the entries in each row of matrix A.
-     *                                Note: vector_sum_row can be a nullptr in case b_offset = 0. Data type supported: same as @p mm_result
-     * @param[in]      bias           Biases tensor. Only shared biases supported and it can be a nullptr if the addition of biases is not required.
-     *                                Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p input.
-     * @param[in]      k              Number of matrix A columns or Matrix B rows
-     * @param[in]      a_offset       Offset to be added to each element of the matrix A.
-     * @param[in]      b_offset       Offset to be added to each element of the matrix B.
-     */
-    void configure(ICLTensor *mm_result, const ICLTensor *vector_sum_col, const ICLTensor *vector_sum_row, const ICLTensor *bias, int32_t k, int32_t a_offset, int32_t b_offset);
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in]      compile_context The compile context to be used.
-     * @param[in, out] mm_result       Input tensor containing the result of the matrix multiplication. Data type supported: S32
-     * @param[in]      vector_sum_col  Input row-vector of sums of all the entries in each column of matrix B.
-     *                                 Note: vector_sum_col can be a nullptr in case a_offset = 0. Data type supported: same as @p mm_result
-     * @param[in]      vector_sum_row  Input row-vector of sums of all the entries in each row of matrix A.
-     *                                 Note: vector_sum_row can be a nullptr in case b_offset = 0. Data type supported: same as @p mm_result
-     * @param[in]      bias            Biases tensor. Only shared biases supported and it can be a nullptr if the addition of biases is not required.
-     *                                 Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p input.
-     * @param[in]      k               Number of matrix A columns or Matrix B rows
-     * @param[in]      a_offset        Offset to be added to each element of the matrix A.
-     * @param[in]      b_offset        Offset to be added to each element of the matrix B.
-     */
-    void configure(const CLCompileContext &compile_context, ICLTensor *mm_result, const ICLTensor *vector_sum_col, const ICLTensor *vector_sum_row, const ICLTensor *bias, int32_t k, int32_t a_offset,
-                   int32_t b_offset);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLGEMMLowpOffsetContributionKernel
-     *
-     * @param[in] mm_result      Input tensor containing the result of @ref CLGEMMLowpOffsetContributionKernel. Data type supported: S32
-     * @param[in] vector_sum_col Input row-vector of sums of all the entries in each column of matrix B.
-     *                           Note: vector_sum_col can be a nullptr in case a_offset = 0. Data type supported: same as @p mm_result
-     * @param[in] vector_sum_row Input row-vector of sums of all the entries in each row of matrix A.
-     *                           Note: vector_sum_row can be a nullptr in case b_offset = 0. Data type supported: same as @p mm_result
-     * @param[in] bias           Biases tensor. Only shared biases supported and it can be a nullptr if the addition of biases is not required.
-     *                           Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p input.
-     * @param[in] a_offset       Offset to be added to each element of the matrix A.
-     * @param[in] b_offset       Offset to be added to each element of the matrix B.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *mm_result, const ITensorInfo *vector_sum_col, const ITensorInfo *vector_sum_row, const ITensorInfo *bias, int32_t a_offset, int32_t b_offset);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_vector_sum_col;
-    const ICLTensor *_vector_sum_row;
-    ICLTensor       *_mm_result;
-    const ICLTensor *_bias;
-};
-} // namespace arm_compute
-
-#endif /* ARM_COMPUTE_CLGEMMLOWPOFFSETCONTRIBUTIONKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLGEMMLowpOffsetContributionOutputStageKernel.h b/arm_compute/core/CL/kernels/CLGEMMLowpOffsetContributionOutputStageKernel.h
deleted file mode 100644
index e3f88c11e6..0000000000
--- a/arm_compute/core/CL/kernels/CLGEMMLowpOffsetContributionOutputStageKernel.h
+++ /dev/null
@@ -1,135 +0,0 @@
-/*
- * Copyright (c) 2018-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLGEMMLOWPOFFSETCONTRIBUTIONOUTPUTSTAGEKERNEL_H
-#define ARM_COMPUTE_CLGEMMLOWPOFFSETCONTRIBUTIONOUTPUTSTAGEKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** OpenCL kernel used to add the offset contribution after the matrix multiplication and perform the output stage.
- *
- * This kernel takes a final int32 accumulator value (the output of the matrix multiplication), adds to it the offset contribution
- * of matrix A and matrix B and performs the output stage defined by the output_stage argument
- *
- * @note For quantized computations the output data type for auto-initialization must be passed as part of the @ref GEMMLowpOutputStageInfo.
- */
-class CLGEMMLowpOffsetContributionOutputStageKernel : public ICLKernel
-{
-public:
-    /** Constructor */
-    CLGEMMLowpOffsetContributionOutputStageKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers)*/
-    CLGEMMLowpOffsetContributionOutputStageKernel(const CLGEMMLowpOffsetContributionOutputStageKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers)*/
-    CLGEMMLowpOffsetContributionOutputStageKernel &operator=(const CLGEMMLowpOffsetContributionOutputStageKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLGEMMLowpOffsetContributionOutputStageKernel(CLGEMMLowpOffsetContributionOutputStageKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLGEMMLowpOffsetContributionOutputStageKernel &operator=(CLGEMMLowpOffsetContributionOutputStageKernel &&) = default;
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in]  mm_result          Input tensor containing the result of the matrix multiplication. Data type supported: S32
-     * @param[in]  vector_sum_col     Input row-vector of sums of all the entries in each column of matrix B.
-     *                                Note: vector_sum_col can be a nullptr in case a_offset = 0. Data type supported: same as @p mm_result
-     * @param[in]  vector_sum_row     Input row-vector of sums of all the entries in each row of matrix A.
-     *                                Note: vector_sum_row can be a nullptr in case b_offset = 0. Data type supported: same as @p mm_result
-     * @param[in]  bias               Biases tensor. Only shared biases supported and it can be a nullptr if the addition of biases is not required.
-     *                                Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p input.
-     * @param[out] output             Output tensor. Data type supported: QASYMM8/QASYMM8_SIGNED.
-     * @param[in]  k                  Number of matrix A columns or Matrix B rows
-     * @param[in]  a_offset           Offset to be added to each element of the matrix A.
-     * @param[in]  b_offset           Offset to be added to each element of the matrix B.
-     * @param[in]  output_stage       GEMMLowp output stage info
-     * @param[in]  output_multipliers Output multipliers tensor. In case of per-channel quantization, the number of multipliers must be equal to the number of filters (OFM).
-     *                                Supported data types: S32
-     * @param[in]  output_shifts      Output shifts tensor. In case of per-channel quantization, the number of multipliers must be equal to the number of filters (OFM).
-     *                                Supported data types: S32
-     */
-    void configure(const ICLTensor *mm_result, const ICLTensor *vector_sum_col, const ICLTensor *vector_sum_row, const ICLTensor *bias, ICLTensor *output, int32_t k, int32_t a_offset, int32_t b_offset,
-                   const GEMMLowpOutputStageInfo &output_stage, const ICLTensor *output_multipliers, const ICLTensor *output_shifts);
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in]  compile_context    The compile context to be used.
-     * @param[in]  mm_result          Input tensor containing the result of the matrix multiplication. Data type supported: S32
-     * @param[in]  vector_sum_col     Input row-vector of sums of all the entries in each column of matrix B.
-     *                                Note: vector_sum_col can be a nullptr in case a_offset = 0. Data type supported: same as @p mm_result
-     * @param[in]  vector_sum_row     Input row-vector of sums of all the entries in each row of matrix A.
-     *                                Note: vector_sum_row can be a nullptr in case b_offset = 0. Data type supported: same as @p mm_result
-     * @param[in]  bias               Biases tensor. Only shared biases supported and it can be a nullptr if the addition of biases is not required.
-     *                                Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p input.
-     * @param[out] output             Output tensor. Data type supported: QASYMM8/QASYMM8_SIGNED.
-     * @param[in]  k                  Number of matrix A columns or Matrix B rows
-     * @param[in]  a_offset           Offset to be added to each element of the matrix A.
-     * @param[in]  b_offset           Offset to be added to each element of the matrix B.
-     * @param[in]  output_stage       GEMMLowp output stage info
-     * @param[in]  output_multipliers Output multipliers tensor. In case of per-channel quantization, the number of multipliers must be equal to the number of filters (OFM).
-     *                                Supported data types: S32
-     * @param[in]  output_shifts      Output shifts tensor. In case of per-channel quantization, the number of multipliers must be equal to the number of filters (OFM).
-     *                                Supported data types: S32
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *mm_result, const ICLTensor *vector_sum_col, const ICLTensor *vector_sum_row, const ICLTensor *bias, ICLTensor *output, int32_t k,
-                   int32_t a_offset, int32_t b_offset,
-                   const GEMMLowpOutputStageInfo &output_stage, const ICLTensor *output_multipliers, const ICLTensor *output_shifts);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLGEMMLowpOffsetContributionKernel
-     *
-     * @param[in] mm_result          Input tensor containing the result of @ref CLGEMMLowpOffsetContributionKernel. Data type supported: S32
-     * @param[in] vector_sum_col     Input row-vector of sums of all the entries in each column of matrix B.
-     *                               Note: vector_sum_col can be a nullptr in case a_offset = 0. Data type supported: same as @p mm_result
-     * @param[in] vector_sum_row     Input row-vector of sums of all the entries in each row of matrix A.
-     *                               Note: vector_sum_row can be a nullptr in case b_offset = 0. Data type supported: same as @p mm_result
-     * @param[in] bias               Biases tensor. Only shared biases supported and it can be a nullptr if the addition of biases is not required.
-     *                               Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p input.
-     * @param[in] output             Output tensor. Data type supported: QASYMM8/QASYMM8_SIGNED.
-     * @param[in] a_offset           Offset to be added to each element of the matrix A.
-     * @param[in] b_offset           Offset to be added to each element of the matrix B.
-     * @param[in] output_stage       GEMMLowp output stage info
-     * @param[in] output_multipliers Output multipliers tensor info. In case of per-channel quantization, the number of multipliers must be equal to the number of filters (OFM).
-     *                               Supported data types: S32
-     * @param[in] output_shifts      Output shifts tensor info. In case of per-channel quantization, the number of multipliers must be equal to the number of filters (OFM).
-     *                               Supported data types: S32
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *mm_result, const ITensorInfo *vector_sum_col, const ITensorInfo *vector_sum_row, const ITensorInfo *bias, const ITensorInfo *output, int32_t a_offset,
-                           int32_t b_offset, const GEMMLowpOutputStageInfo &output_stage, const ITensorInfo *output_multipliers, const ITensorInfo *output_shifts);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_mm_result;
-    const ICLTensor *_vector_sum_col;
-    const ICLTensor *_vector_sum_row;
-    const ICLTensor *_bias;
-    ICLTensor       *_output;
-    const ICLTensor *_output_multipliers;
-    const ICLTensor *_output_shifts;
-    bool             _is_quantized_per_channel;
-};
-} // namespace arm_compute
-
-#endif /* ARM_COMPUTE_CLGEMMLOWPOFFSETCONTRIBUTIONOUTPUTSTAGEKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLGEMMLowpQuantizeDownInt32ScaleByFixedPointKernel.h b/arm_compute/core/CL/kernels/CLGEMMLowpQuantizeDownInt32ScaleByFixedPointKernel.h
deleted file mode 100644
index 23040e7bcc..0000000000
--- a/arm_compute/core/CL/kernels/CLGEMMLowpQuantizeDownInt32ScaleByFixedPointKernel.h
+++ /dev/null
@@ -1,89 +0,0 @@
-/*
- * Copyright (c) 2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLGEMMLOWPQUANTIZEDOWNINT32SCALEBYFIXEDPOINTKERNEL_H
-#define ARM_COMPUTE_CLGEMMLOWPQUANTIZEDOWNINT32SCALEBYFIXEDPOINTKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** OpenCL kernel used to quantize down the int32 accumulator values of GEMMLowp to QASYMM8/QASYMM8_SIGNED/QSYMM16
- *
- * This kernel takes a final int32 accumulator value (the output of the matrix multiplication), and processes it to obtain the final quantized value.
- * The following computations will be performed by the kernel:
- *
- *  -# Compute fixed point multiplication between each entry of input by gemmlowp_multiplier
- *  -# Add bias to final result if bias tensor is not a nullptr
- *  -# Round to nearest division by a power-of-two using result_shift
- *  -# Add offset to each result
- *  -# Clamp the value between the specified min and max bounds
- *  -# Clamp the resulting int32 values to the proper quantized range and cast to QASYMM8/QASYMM8_SIGNED/QSYMM16.
- */
-class CLGEMMLowpQuantizeDownInt32ScaleByFixedPointKernel : public ICLKernel
-{
-public:
-    /** Constructor */
-    CLGEMMLowpQuantizeDownInt32ScaleByFixedPointKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers)*/
-    CLGEMMLowpQuantizeDownInt32ScaleByFixedPointKernel(const CLGEMMLowpQuantizeDownInt32ScaleByFixedPointKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers)*/
-    CLGEMMLowpQuantizeDownInt32ScaleByFixedPointKernel &operator=(const CLGEMMLowpQuantizeDownInt32ScaleByFixedPointKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLGEMMLowpQuantizeDownInt32ScaleByFixedPointKernel(CLGEMMLowpQuantizeDownInt32ScaleByFixedPointKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLGEMMLowpQuantizeDownInt32ScaleByFixedPointKernel &operator=(CLGEMMLowpQuantizeDownInt32ScaleByFixedPointKernel &&) = default;
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Input tensor. Data type supported: S32
-     * @param[in]  bias            Biases tensor. Only shared biases supported and it can be a nullptr if the biases addition is not required.
-     *                             Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p input.
-     * @param[out] output          Output tensor. Data type supported: Data type supported: QASYMM8/QASYMM8_SIGNED/QSYMM16.
-     * @param[in]  info            Output stage info. Used to pass the quantized output data type
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, const ICLTensor *bias, ICLTensor *output, const GEMMLowpOutputStageInfo *info);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLGEMMLowpQuantizeDownInt32ScaleByFixedPointKernel
-     *
-     * @param[in] input  Input tensor. Data type supported: S32
-     * @param[in] bias   Biases tensor. Only shared biases supported and it can be a nullptr if the biases addition is not required.
-     *                   Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p input.
-     * @param[in] output Output tensor. Data type supported: Data type supported: QSYMM8/QASYMM8_SIGNED/QSYMM16.
-     * @param[in] info   Output stage info. Used to pass the quantized output data type
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *bias, const ITensorInfo *output, const GEMMLowpOutputStageInfo *info);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;
-    const ICLTensor *_bias;
-    ICLTensor       *_output;
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLGEMMLOWPQUANTIZEDOWNINT32SCALEBYFIXEDPOINTKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLGEMMLowpQuantizeDownInt32ScaleByFloatKernel.h b/arm_compute/core/CL/kernels/CLGEMMLowpQuantizeDownInt32ScaleByFloatKernel.h
deleted file mode 100644
index 0b3f23dab3..0000000000
--- a/arm_compute/core/CL/kernels/CLGEMMLowpQuantizeDownInt32ScaleByFloatKernel.h
+++ /dev/null
@@ -1,101 +0,0 @@
-/*
- * Copyright (c) 2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLGEMMLOWPQUANTIZEDOWNINT32SCALEBYFLOATKERNEL_H
-#define ARM_COMPUTE_CLGEMMLOWPQUANTIZEDOWNINT32SCALEBYFLOATKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-// Forward declarations
-class ICLTensor;
-
-/** OpenCL kernel used to quantize down the int32 accumulator values of GEMMLowp to QASYMM8/QASYMM8_SIGNED
- *
- * This kernel takes a final int32 accumulator value (the output of the matrix multiplication), and processes it to obtain the final QASYMM8/QASYMM8_SIGNED value.
- * The following computations will be performed by the kernel:
- *
- *  -# Compute fixed point multiplication between each entry of input by result_fixedpoint_multiplier
- *  -# Add bias to final result if bias tensor is not a nullptr
- *  -# Requantize
- *  -# Add offset to each result
- *  -# Clamp the value between the specified min and max bounds
- *  -# Clamp the resulting int32 values to
- *      - to the [0..255] range and cast to QASYMM8.
- *      - to the [-128..127] range and cast to QASYMM8_SIGNED.
- */
-class CLGEMMLowpQuantizeDownInt32ScaleByFloatKernel : public ICLKernel
-{
-public:
-    /** Constructor */
-    CLGEMMLowpQuantizeDownInt32ScaleByFloatKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers)*/
-    CLGEMMLowpQuantizeDownInt32ScaleByFloatKernel(const CLGEMMLowpQuantizeDownInt32ScaleByFloatKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers)*/
-    CLGEMMLowpQuantizeDownInt32ScaleByFloatKernel &operator=(const CLGEMMLowpQuantizeDownInt32ScaleByFloatKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLGEMMLowpQuantizeDownInt32ScaleByFloatKernel(CLGEMMLowpQuantizeDownInt32ScaleByFloatKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLGEMMLowpQuantizeDownInt32ScaleByFloatKernel &operator=(CLGEMMLowpQuantizeDownInt32ScaleByFloatKernel &&) = default;
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in]  input  Input tensor. Data type supported: S32
-     * @param[in]  bias   Biases tensor. Only shared biases supported and it can be a nullptr if the biases addition is not required.
-     *                    Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p input.
-     * @param[out] output Output tensor. Data type supported: Data type supported: QASYMM8/QASYMM8_SIGNED
-     * @param[in]  info   Output stage info. Used to pass the quantized output data type
-     */
-    void configure(const ICLTensor *input, const ICLTensor *bias, ICLTensor *output, const GEMMLowpOutputStageInfo *info);
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Input tensor. Data type supported: S32
-     * @param[in]  bias            Biases tensor. Only shared biases supported and it can be a nullptr if the biases addition is not required.
-     *                             Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p input.
-     * @param[out] output          Output tensor. Data type supported: Data type supported: QASYMM8/QASYMM8_SIGNED
-     * @param[in]  info            Output stage info. Used to pass the quantized output data type
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, const ICLTensor *bias, ICLTensor *output, const GEMMLowpOutputStageInfo *info);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLGEMMLowpQuantizeDownInt32ScaleByFloatKernel
-     *
-     * @param[in] input  Input tensor. Data type supported: S32
-     * @param[in] bias   Biases tensor. Only shared biases supported and it can be a nullptr if the biases addition is not required.
-     *                   Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p input.
-     * @param[in] output Output tensor. Data type supported: Data type supported: QASYMM8/QASYMM8_SIGNED
-     * @param[in] info   Output stage info. Used to pass the quantized output data type
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *bias, const ITensorInfo *output, const GEMMLowpOutputStageInfo *info);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;
-    const ICLTensor *_bias;
-    ICLTensor       *_output;
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLGEMMLOWPQUANTIZEDOWNINT32SCALEBYFLOATKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLGEMMLowpQuantizeDownInt32ScaleKernel.h b/arm_compute/core/CL/kernels/CLGEMMLowpQuantizeDownInt32ScaleKernel.h
deleted file mode 100644
index 1a284f0701..0000000000
--- a/arm_compute/core/CL/kernels/CLGEMMLowpQuantizeDownInt32ScaleKernel.h
+++ /dev/null
@@ -1,102 +0,0 @@
-/*
- * Copyright (c) 2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLGEMMLOWPQUANTIZEDOWNINT32SCALEKERNEL_H
-#define ARM_COMPUTE_CLGEMMLOWPQUANTIZEDOWNINT32SCALEKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** OpenCL kernel used to quantize down the int32 accumulator values of GEMMLowp to QASYMM8/QASYMM8_SIGNED
- *
- * This kernel takes a final int32 accumulator value (the output of the matrix multiplication), and processes it to obtain the final QASYMM8/QASYMM8_SIGNED value.
- * The following computations will be performed by the kernel:
- *
- *  -# Add offset terms to final result
- *  -# Multiply each entry of result by result_mult_int
- *  -# Add bias to final result if bias tensor is not a nullptr
- *  -# Shift the int32 accumulator by result_shift
- *  -# Clamp the value between the specified min and max bounds
- *  -# Clamp the resulting int32 values:
- *  -#  -to the [0..255] range and cast to QASYMM8.
- *  -#  -to the [-128..127] range and cast to QASYMM8_SIGNED.
- *
- */
-class CLGEMMLowpQuantizeDownInt32ScaleKernel : public ICLKernel
-{
-public:
-    /** Constructor */
-    CLGEMMLowpQuantizeDownInt32ScaleKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers)*/
-    CLGEMMLowpQuantizeDownInt32ScaleKernel(const CLGEMMLowpQuantizeDownInt32ScaleKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers)*/
-    CLGEMMLowpQuantizeDownInt32ScaleKernel &operator=(const CLGEMMLowpQuantizeDownInt32ScaleKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLGEMMLowpQuantizeDownInt32ScaleKernel(CLGEMMLowpQuantizeDownInt32ScaleKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLGEMMLowpQuantizeDownInt32ScaleKernel &operator=(CLGEMMLowpQuantizeDownInt32ScaleKernel &&) = default;
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in]  input        Input tensor. Data type supported: S32
-     * @param[in]  bias         Biases tensor. Only shared biases supported and it can be a nullptr if the biases addition is not required.
-     *                          Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p input.
-     * @param[out] output       Output tensor. Data type supported: Data type supported: QASYMM8/QASYMM8_SIGNED
-     * @param[in]  output_stage GEMMLowp output stage metadata.
-     */
-    void configure(const ICLTensor *input, const ICLTensor *bias, ICLTensor *output, const GEMMLowpOutputStageInfo *output_stage);
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Input tensor. Data type supported: S32
-     * @param[in]  bias            Biases tensor. Only shared biases supported and it can be a nullptr if the biases addition is not required.
-     *                             Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p input.
-     * @param[out] output          Output tensor. Data type supported: Data type supported: QASYMM8/QASYMM8_SIGNED
-     * @param[in]  output_stage    GEMMLowp output stage metadata.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, const ICLTensor *bias, ICLTensor *output, const GEMMLowpOutputStageInfo *output_stage);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLGEMMLowpQuantizeDownInt32ScaleKernel
-     *
-     * @param[in] input        Input tensor. Data type supported: S32
-     * @param[in] bias         Biases tensor. Only shared biases supported and it can be a nullptr if the biases addition is not required.
-     *                         Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p input.
-     * @param[in] output       Output tensor. Data type supported: Data type supported: QASYMM8/QASYMM8_SIGNED
-     * @param[in] output_stage GEMMLowp output stage metadata.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *bias, const ITensorInfo *output, const GEMMLowpOutputStageInfo *output_stage);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;
-    const ICLTensor *_bias;
-    ICLTensor       *_output;
-};
-} // namespace arm_compute
-
-#endif /* ARM_COMPUTE_CLGEMMLOWPQUANTIZEDOWNINT32SCALEKERNEL_H */
\ No newline at end of file
diff --git a/arm_compute/core/CL/kernels/CLGEMMLowpReductionKernel.h b/arm_compute/core/CL/kernels/CLGEMMLowpReductionKernel.h
deleted file mode 100644
index 6066e2a815..0000000000
--- a/arm_compute/core/CL/kernels/CLGEMMLowpReductionKernel.h
+++ /dev/null
@@ -1,176 +0,0 @@
-/*
- * Copyright (c) 2017-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLGEMMLOWREDUCTIONKERNEL_H
-#define ARM_COMPUTE_CLGEMMLOWREDUCTIONKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-struct GEMMLowpReductionKernelInfo;
-
-/** Common interface for all OpenCL reduction kernels */
-class ICLGEMMLowpReductionKernel : public ICLKernel
-{
-public:
-    /** Constructor */
-    ICLGEMMLowpReductionKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers)*/
-    ICLGEMMLowpReductionKernel(const ICLGEMMLowpReductionKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers)*/
-    ICLGEMMLowpReductionKernel &operator=(const ICLGEMMLowpReductionKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    ICLGEMMLowpReductionKernel(ICLGEMMLowpReductionKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    ICLGEMMLowpReductionKernel &operator=(ICLGEMMLowpReductionKernel &&) = default;
-
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in]  input  Input tensor. Data type supported: QASYMM8/QASYMM8_SIGNED/QSYMM8.
-     * @param[out] output Output row-vector of sums of all the entries in each row/col of input tensor. Data type supported: S32
-     * @param[in]  info   Kernel metadata:
-     *                    - k            Number of matrix columns/rows depending on the type of reduction.
-     *                    - is_reshaped  True if the matrix has been reshaped.
-     *                    - scalar       Scalar value to multiply each reduced column/row by.
-     *                    - mul_byscalar True if each reduced column/row must be multiplied by a scalar value.
-     */
-    virtual void configure(const ICLTensor *input, ICLTensor *output, const GEMMLowpReductionKernelInfo &info) = 0;
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Input tensor. Data type supported: QASYMM8/QASYMM8_SIGNED/QSYMM8.
-     * @param[out] output          Output row-vector of sums of all the entries in each row/col of input tensor. Data type supported: S32
-     * @param[in]  info            Kernel metadata:
-     *                             - k            Number of matrix columns/rows depending on the type of reduction.
-     *                             - is_reshaped  True if the matrix has been reshaped.
-     *                             - scalar       Scalar value to multiply each reduced column/row by.
-     *                             - mul_byscalar True if each reduced column/row must be multiplied by a scalar value.
-     */
-    virtual void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, const GEMMLowpReductionKernelInfo &info) = 0;
-
-protected:
-    const ICLTensor *_input;
-    ICLTensor       *_output;
-};
-
-/** OpenCL kernel used to compute the row-vectors of sums of all the entries in each row of Matrix A.
- *
- * @note This stage is needed to handle the offset of matrix product
- *       https://github.com/google/gemmlowp/blob/master/doc/low-precision.md
- */
-class CLGEMMLowpMatrixAReductionKernel : public ICLGEMMLowpReductionKernel
-{
-public:
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in]  mtx_a          Input tensor. Data type supported: QASYMM8/QASYMM8_SIGNED/QSYMM8.
-     * @param[out] vector_sum_row Output row-vector of sums of all the entries in each row of mtx_a. Data type supported: S32
-     * @param[in]  info           Kernel metadata:
-     *                            - k            Number of matrix columns/rows depending on the type of reduction.
-     *                            - is_reshaped  True if the matrix has been reshaped.
-     *                            - scalar       Scalar value to multiply each reduced column/row by.
-     *                            - mul_byscalar True if each reduced column/row must be multiplied by a scalar value.
-     */
-    void configure(const ICLTensor *mtx_a, ICLTensor *vector_sum_row, const GEMMLowpReductionKernelInfo &info) override;
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  mtx_a           Input tensor. Data type supported: QASYMM8/QASYMM8_SIGNED/QSYMM8.
-     * @param[out] vector_sum_row  Output row-vector of sums of all the entries in each row of mtx_a. Data type supported: S32
-     * @param[in]  info            Kernel metadata:
-     *                             - k            Number of matrix columns/rows depending on the type of reduction.
-     *                             - is_reshaped  True if the matrix has been reshaped.
-     *                             - scalar       Scalar value to multiply each reduced column/row by.
-     *                             - mul_byscalar True if each reduced column/row must be multiplied by a scalar value.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *mtx_a, ICLTensor *vector_sum_row, const GEMMLowpReductionKernelInfo &info) override;
-    /** Static function to check if given info will lead to a valid configuration of @ref CLGEMMLowpMatrixAReductionKernel
-     *
-     * @param[in] mtx_a          Input tensor. Data type supported: QASYMM8/QASYMM8_SIGNED/QSYMM8.
-     * @param[in] vector_sum_row Output row-vector of sums of all the entries in each row of mtx_a. Data type supported: S32
-     * @param[in] info           Kernel metadata:
-     *                           - k            Number of matrix columns/rows depending on the type of reduction.
-     *                           - is_reshaped  True if the matrix has been reshaped.
-     *                           - scalar       Scalar value to multiply each reduced column/row by.
-     *                           - mul_byscalar True if each reduced column/row must be multiplied by a scalar value.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *mtx_a, const ITensorInfo *vector_sum_row, const GEMMLowpReductionKernelInfo &info);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-};
-
-/** OpenCL kernel used to compute the row-vectors of sums of all the entries in each column of Matrix B.
- *
- * @note This stage is needed to handle the offset of matrix product
- *       https://github.com/google/gemmlowp/blob/master/doc/low-precision.md
- */
-class CLGEMMLowpMatrixBReductionKernel : public ICLGEMMLowpReductionKernel
-{
-public:
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in]  mtx_b          Input tensor. Data type supported: Data type supported: QASYMM8/QASYMM8_SIGNED/QSYMM8/QSYMM8_PER_CHANNEL.
-     * @param[out] vector_sum_col Output row-vector of sums of all the entries in each column of mtx_b. Data type supported: S32
-     * @param[in]  info           Kernel metadata:
-     *                            - k            Number of matrix columns/rows depending on the type of reduction.
-     *                            - is_reshaped  True if the matrix has been reshaped.
-     *                            - scalar       Scalar value to multiply each reduced column/row by.
-     *                            - mul_byscalar True if each reduced column/row must be multiplied by a scalar value.
-     */
-    void configure(const ICLTensor *mtx_b, ICLTensor *vector_sum_col, const GEMMLowpReductionKernelInfo &info) override;
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  mtx_b           Input tensor. Data type supported: Data type supported: QASYMM8/QASYMM8_SIGNED/QSYMM8/QSYMM8_PER_CHANNEL.
-     * @param[out] vector_sum_col  Output row-vector of sums of all the entries in each column of mtx_b. Data type supported: S32
-     * @param[in]  info            Kernel metadata:
-     *                             - k            Number of matrix columns/rows depending on the type of reduction.
-     *                             - is_reshaped  True if the matrix has been reshaped.
-     *                             - scalar       Scalar value to multiply each reduced column/row by.
-     *                             - mul_byscalar True if each reduced column/row must be multiplied by a scalar value.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *mtx_b, ICLTensor *vector_sum_col, const GEMMLowpReductionKernelInfo &info) override;
-    /** Static function to check if given info will lead to a valid configuration of @ref CLGEMMLowpMatrixBReductionKernel
-     *
-     * @param[in] mtx_b          Input tensor. Data type supported: Data type supported: QASYMM8/QASYMM8_SIGNED/QSYMM8/QSYMM8_PER_CHANNEL.
-     * @param[in] vector_sum_col Output row-vector of sums of all the entries in each column of mtx_b. Data type supported: S32
-     * @param[in] info           Kernel metadata:
-     *                           - k            Number of matrix columns/rows depending on the type of reduction.
-     *                           - is_reshaped  True if the matrix has been reshaped.
-     *                           - scalar       Scalar value to multiply each reduced column/row by.
-     *                           - mul_byscalar True if each reduced column/row must be multiplied by a scalar value.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *mtx_b, const ITensorInfo *vector_sum_col, const GEMMLowpReductionKernelInfo &info);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLGEMMLOWREDUCTIONKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLGEMMMatrixMultiplyKernel.h b/arm_compute/core/CL/kernels/CLGEMMMatrixMultiplyKernel.h
deleted file mode 100644
index 4abd60c202..0000000000
--- a/arm_compute/core/CL/kernels/CLGEMMMatrixMultiplyKernel.h
+++ /dev/null
@@ -1,122 +0,0 @@
-/*
- * Copyright (c) 2017-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLGEMMMATRIXMULTIPLYKERNEL_H
-#define ARM_COMPUTE_CLGEMMMATRIXMULTIPLYKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** OpenCL kernel to multiply two input matrices "A" and "B" and add a martix "C" if provided. All elements of the output matrix will be multiplied by alpha. In case matrix C is passed, it will be added to the previous result.
- *  For the matrix C, the broadcast addition is supported if the flag "broadcast_bias" is set in the GEMMReshapeInfo object
- *
- * @note If the input tensors @p input0 and @p input1 have been reshaped respectively with @ref CLGEMMReshapeLHSMatrixKernel" and @ref CLGEMMReshapeRHSMatrixKernel,
- *       the flag @p is_interleaved_transposed must be set to true
- *
- * @attention @p input1 tensor must have at least 2 dimensions (matrix)
- *
- */
-class CLGEMMMatrixMultiplyKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLGEMMMatrixMultiplyKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLGEMMMatrixMultiplyKernel(const CLGEMMMatrixMultiplyKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLGEMMMatrixMultiplyKernel &operator=(const CLGEMMMatrixMultiplyKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLGEMMMatrixMultiplyKernel(CLGEMMMatrixMultiplyKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLGEMMMatrixMultiplyKernel &operator=(CLGEMMMatrixMultiplyKernel &&) = default;
-    /** Initialise the kernel's input, output and alpha
-     *
-     * @param[in]  input0                    Input tensor containing the Matrix A. Data types supported: F16/F32
-     * @param[in]  input1                    Input tensor containing the Matrix B. Data type supported: same as @p input0
-     * @param[in]  input2                    Input tensor containing the Matrix C (bias). Can be nullptr. Data type supported: same as @p input0
-     * @param[out] output                    Output tensor to store the result of matrix multiplication. Data type supported: same as @p input0
-     * @param[in]  alpha                     Weight of the matrix product
-     * @param[in]  beta                      (Optional) Weight of vector C. Default value is 0. Only beta = 1 is currently supported.
-     * @param[in]  is_interleaved_transposed (Optional) True if input0 and input1 have been reshaped respectively using @ref CLGEMMReshapeLHSMatrixKernel and @ref CLGEMMReshapeRHSMatrixKernel
-     * @param[in]  reshape_info              (Optional) GEMM reshape info. If is_interleaved_transposed = true, this object must contain the information to understand how the matrix A and matrix B have been reshaped
-     * @param[in]  fp_mixed_precision        (Optional) Use wider accumulators (32 bit instead of 16 for FP16) to improve accuracy
-     * @param[in]  activation_info           (Optional) Activation to apply after the matrix multiplication
-     *
-     */
-    void configure(const ICLTensor *input0, const ICLTensor *input1, const ICLTensor *input2, ICLTensor *output, float alpha, float beta = 0.f,
-                   bool is_interleaved_transposed = true, const GEMMReshapeInfo &reshape_info = GEMMReshapeInfo(), bool fp_mixed_precision = false, const ActivationLayerInfo &activation_info = ActivationLayerInfo());
-    /** Initialise the kernel's input, output and alpha
-     *
-     * @param[in]  compile_context           The compile context to be used.
-     * @param[in]  input0                    Input tensor containing the Matrix A. Data types supported: F16/F32
-     * @param[in]  input1                    Input tensor containing the Matrix B. Data type supported: same as @p input0
-     * @param[in]  input2                    Input tensor containing the Matrix C (bias). Can be nullptr. Data type supported: same as @p input0
-     * @param[out] output                    Output tensor to store the result of matrix multiplication. Data type supported: same as @p input0
-     * @param[in]  alpha                     Weight of the matrix product
-     * @param[in]  beta                      (Optional) Weight of vector C. Default value is 0. Only beta = 1 is currently supported.
-     * @param[in]  is_interleaved_transposed (Optional) True if input0 and input1 have been reshaped respectively using @ref CLGEMMReshapeLHSMatrixKernel and @ref CLGEMMReshapeRHSMatrixKernel
-     * @param[in]  reshape_info              (Optional) GEMM reshape info. If is_interleaved_transposed = true, this object must contain the information to understand how the matrix A and matrix B have been reshaped
-     * @param[in]  fp_mixed_precision        (Optional) Use wider accumulators (32 bit instead of 16 for FP16) to improve accuracy
-     * @param[in]  activation_info           (Optional) Activation to apply after the matrix multiplication
-     *
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input0, const ICLTensor *input1, const ICLTensor *input2, ICLTensor *output, float alpha, float beta = 0.f,
-                   bool is_interleaved_transposed = true, const GEMMReshapeInfo &reshape_info = GEMMReshapeInfo(), bool fp_mixed_precision = false, const ActivationLayerInfo &activation_info = ActivationLayerInfo());
-    /** Static function to check if given info will lead to a valid configuration of @ref CLGEMMMatrixMultiplyKernel
-     *
-     * @param[in] input0                    Input tensor containing the Matrix A info. Data types supported: F16/F32
-     * @param[in] input1                    Input tensor containing the Matrix B info. Data type supported: same as @p input0
-     * @param[in] input2                    Input tensor containing the Matrix C (bias) info. Can be nullptr. Data type supported: same as @p input0
-     * @param[in] output                    Output tensor to store the result of matrix multiplication. Data type supported: same as @p input0
-     * @param[in] alpha                     Weight of the matrix product
-     * @param[in] beta                      Weight of vector C. Default value is 0. Only beta = 1 is currently supported.
-     * @param[in] is_interleaved_transposed True if input0 and input1 have been reshaped respectively using @ref CLGEMMReshapeLHSMatrixKernel and @ref CLGEMMReshapeRHSMatrixKernel
-     * @param[in] reshape_info              GEMM reshape info. If is_interleaved_transposed = true, this object must contain the information to understand how the matrix A and matrix B have been reshaped
-     * @param[in] gpu_target                GPU Target
-     * @param[in] fp_mixed_precision        (Optional) Use wider accumulators (32 bit instead of 16 for FP16) to improve accuracy
-     * @param[in] activation_info           (Optional) Activation to apply after the matrix multiplication
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input0, const ITensorInfo *input1, const ITensorInfo *input2, const ITensorInfo *output, float alpha, float beta,
-                           bool is_interleaved_transposed, const GEMMReshapeInfo &reshape_info, GPUTarget gpu_target, bool fp_mixed_precision = false, const ActivationLayerInfo &activation_info = ActivationLayerInfo());
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-public:
-    const ICLTensor *_input0;
-    const ICLTensor *_input1;
-    const ICLTensor *_input2;
-    ICLTensor       *_output;
-    bool             _slide_matrix_b;
-    bool             _reinterpret_input_as_3d;
-    bool             _reinterpret_output_as_3d;
-    bool             _add_bias;
-    bool             _broadcast_bias;
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLGEMMMATRIXMULTIPLYKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLGEMMMatrixMultiplyNativeKernel.h b/arm_compute/core/CL/kernels/CLGEMMMatrixMultiplyNativeKernel.h
deleted file mode 100644
index 006b2bf91f..0000000000
--- a/arm_compute/core/CL/kernels/CLGEMMMatrixMultiplyNativeKernel.h
+++ /dev/null
@@ -1,127 +0,0 @@
-/*
- * Copyright (c) 2019-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLGEMMMATRIXMULTIPLYNATIVEKERNEL_H
-#define ARM_COMPUTE_CLGEMMMATRIXMULTIPLYNATIVEKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-#include "arm_compute/core/KernelDescriptors.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** OpenCL kernel to multiply matrices when neither of the input matrices have been reshaped */
-class CLGEMMMatrixMultiplyNativeKernel : public ICLKernel
-{
-public:
-    /** Default Constructor */
-    CLGEMMMatrixMultiplyNativeKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLGEMMMatrixMultiplyNativeKernel(const CLGEMMMatrixMultiplyNativeKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLGEMMMatrixMultiplyNativeKernel &operator=(const CLGEMMMatrixMultiplyNativeKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLGEMMMatrixMultiplyNativeKernel(CLGEMMMatrixMultiplyNativeKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLGEMMMatrixMultiplyNativeKernel &operator=(CLGEMMMatrixMultiplyNativeKernel &&) = default;
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in]  input0    Input tensor for the LHS matrix. Data type supported: F32. The number of dimensions for the LHS matrix must be less or equal than 4.
-     * @param[in]  input1    Input tensor for the RHS matrix. Data type supported: same as @p input0. The number of dimensions for the RHS matrix must be less or equal than 3.
-     * @param[in]  input2    Input tensor containing the bias matrix. Data type supported: same as @p input0.
-     * @param[out] output    Output tensor info. Data type supported: same as @p input0
-     * @param[in]  alpha     Weight of the matrix product
-     * @param[in]  beta      Weight of the matrix bias
-     * @param[in]  lhs_info  LHS matrix information used to retrieve the number of rows and accumulations to be processed by each thread. Only the following values are supported:
-     *                       lhs_info.m0: 1,2,3,4,5,6,7,8
-     *                       lhs_info.k0: 2,3,4,8,16
-     * @param[in]  rhs_info  RHS matrix information used to retrieve the number of columns and accumulations to be processed by each thread. Only the following values are supported:
-     *                       rhs_info.n0: 2,3,4,8,16
-     *                       rhs_info.k0: same of lhs_info.k0
-     * @param[in]  gemm_info GEMM information used to retrieve the original dimensions of the input matrices
-     */
-    void configure(const ICLTensor *input0, const ICLTensor *input1, const ICLTensor *input2, ICLTensor *output, float alpha, float beta, const GEMMLHSMatrixInfo &lhs_info,
-                   const GEMMRHSMatrixInfo &rhs_info,
-                   const GEMMKernelInfo    &gemm_info);
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input0          Input tensor for the LHS matrix. Data type supported: F32. The number of dimensions for the LHS matrix must be less or equal than 4.
-     * @param[in]  input1          Input tensor for the RHS matrix. Data type supported: same as @p input0. The number of dimensions for the RHS matrix must be less or equal than 3.
-     * @param[in]  input2          Input tensor containing the bias matrix. Data type supported: same as @p input0.
-     * @param[out] output          Output tensor info. Data type supported: same as @p input0
-     * @param[in]  alpha           Weight of the matrix product
-     * @param[in]  beta            Weight of the matrix bias
-     * @param[in]  lhs_info        LHS matrix information used to retrieve the number of rows and accumulations to be processed by each thread. Only the following values are supported:
-     *                             lhs_info.m0: 1,2,3,4,5,6,7,8
-     *                             lhs_info.k0: 2,3,4,8,16
-     * @param[in]  rhs_info        RHS matrix information used to retrieve the number of columns and accumulations to be processed by each thread. Only the following values are supported:
-     *                             rhs_info.n0: 2,3,4,8,16
-     *                             rhs_info.k0: same of lhs_info.k0
-     * @param[in]  gemm_info       GEMM information used to retrieve the original dimensions of the input matrices
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input0, const ICLTensor *input1, const ICLTensor *input2, ICLTensor *output, float alpha, float beta,
-                   const GEMMLHSMatrixInfo &lhs_info,
-                   const GEMMRHSMatrixInfo &rhs_info,
-                   const GEMMKernelInfo    &gemm_info);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLGEMMMatrixMultiplyNativeKernel
-     *
-     * @param[in] input0    Input tensor info for the LHS matrix. Data type supported: F32. The number of dimensions for the LHS matrix must be less or equal than 4.
-     * @param[in] input1    Input tensor info for the RHS matrix. Data type supported: same as @p input0. The number of dimensions for the RHS matrix must be less or equal than 3.
-     * @param[in] input2    Input tensor info containing the bias matrix. Data type supported: same as @p input0.
-     * @param[in] output    Output tensor info. Data type supported: same as @p input0
-     * @param[in] alpha     Weight of the matrix product
-     * @param[in] beta      Weight of the matrix bias
-     * @param[in] lhs_info  LHS matrix information used to retrieve the number of rows and accumulations to be processed by each thread. Only the following values are supported:
-     *                      lhs_info.m0: 1,2,3,4,5,6,7,8
-     *                      lhs_info.k0: 2,3,4,8,16
-     * @param[in] rhs_info  RHS matrix information used to retrieve the number of columns and accumulations to be processed by each thread. Only the following values are supported:
-     *                      rhs_info.n0: 2,3,4,8,16
-     *                      rhs_info.k0: same of lhs_info.k0
-     * @param[in] gemm_info GEMM information used to retrieve the original dimensions of the input matrices
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input0, const ITensorInfo *input1, const ITensorInfo *input2, const ITensorInfo *output, float alpha, float beta, const GEMMLHSMatrixInfo &lhs_info,
-                           const GEMMRHSMatrixInfo &rhs_info,
-                           const GEMMKernelInfo    &gemm_info);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input0;
-    const ICLTensor *_input1;
-    const ICLTensor *_input2;
-    ICLTensor       *_output;
-    bool             _slide_matrix_b;
-    bool             _reinterpret_input_as_3d;
-    bool             _reinterpret_output_as_3d;
-    bool             _use_dummy_work_items;
-    bool             _add_bias;
-    bool             _broadcast_bias;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLGEMMMATRIXMULTIPLYNATIVEKERNEL_H*/
diff --git a/arm_compute/core/CL/kernels/CLGEMMMatrixMultiplyReshapedKernel.h b/arm_compute/core/CL/kernels/CLGEMMMatrixMultiplyReshapedKernel.h
deleted file mode 100644
index 962645749e..0000000000
--- a/arm_compute/core/CL/kernels/CLGEMMMatrixMultiplyReshapedKernel.h
+++ /dev/null
@@ -1,188 +0,0 @@
-/*
- * Copyright (c) 2018-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLGEMMMATRIXMULTIPLYRESHAPEDKERNEL_H
-#define ARM_COMPUTE_CLGEMMMATRIXMULTIPLYRESHAPEDKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-#include "arm_compute/core/KernelDescriptors.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** OpenCL kernel to multiply matrices when both the input matrices LHS (input0) and RHS (input1) have been reshaped
- *
- * @note The input matrices @p input0 and @p input1 must be reshaped through @ref CLGEMMReshapeLHSMatrixKernel and  @ref CLGEMMReshapeRHSMatrixKernel
- */
-class CLGEMMMatrixMultiplyReshapedKernel : public ICLKernel
-{
-public:
-    /** Default Constructor */
-    CLGEMMMatrixMultiplyReshapedKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLGEMMMatrixMultiplyReshapedKernel(const CLGEMMMatrixMultiplyReshapedKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLGEMMMatrixMultiplyReshapedKernel &operator=(const CLGEMMMatrixMultiplyReshapedKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLGEMMMatrixMultiplyReshapedKernel(CLGEMMMatrixMultiplyReshapedKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLGEMMMatrixMultiplyReshapedKernel &operator=(CLGEMMMatrixMultiplyReshapedKernel &&) = default;
-    /** Initialise the kernel's input and output.
-     *
-     * @note The F16 computation also supports mixed precision through the gemm_info.fp_mixed_precision flag.
-     *       Mixed precision combines different floating precisions during the computation, in particular, F32 for the accumulations and F16 for the
-     *       multiplications. i.e. float c = (half)a * (half)b
-     *
-     * @note If rhs_info.export_to_cl_image = true, this OpenCL kernel will fetch the RHS data using the OpenCL read_image built-in function.
-     *       Reading from the OpenCL image object can increase the performance. However, since the OpenCL image object is created importing the OpenCL buffer,
-     *       the following conditions are required:
-     *       -# rhs_info.n0 can only be 4, 8 and 16
-     *       -# rhs_info.k0 can only be 4, 8 and 16
-     *       -# Data type can only be F32
-     *       -# The platform should support the OpenCL cl_khr_image2d_from_buffer extension
-     *       -# The stride Y for the input1 should satisfy the OpenCL pitch alignment requirement
-     *       -# input1 width should be less or equal to (CL_DEVICE_IMAGE2D_MAX_WIDTH * 4)
-     *       -# input1 (height * depth) should be less or equal to CL_DEVICE_IMAGE2D_MAX_HEIGHT
-     *
-     * @param[in]  input0    Input tensor containing the LHS reshaped matrix. Data type supported: F16/F32 (only F32 if rhs_info.export_to_cl_image = true). The number of dimensions for the LHS matrix must be less or equal than 4
-     * @param[in]  input1    Input tensor containing the RHS reshaped matrix. Data type supported: same as @p input0. The number of dimensions for the RHS matrix must be less or equal than 3
-     * @param[in]  input2    Input tensor containing the bias matrix. Data type supported: same as @p input0.
-     * @param[out] output    Output tensor to store the result of matrix multiplication. Data type supported: same as @p input0
-     * @param[in]  alpha     Weight of the matrix product
-     * @param[in]  beta      Weight of the matrix bias
-     * @param[in]  lhs_info  LHS matrix information used for reshaping the input0 tensor. Only the following values are supported:
-     *                       lhs_info.m0: 2,3,4,5,6,7,8
-     *                       lhs_info.k0: 2,3,4,8,16
-     *                       lhs_info.transpose: false
-     * @param[in]  rhs_info  RHS matrix information used for reshaping the input1 tensor.  Only the following values are supported:
-     *                       rhs_info.n0: 2,3,4,8,16 (only 4, 8 and 16 if rhs_info.export_to_cl_image = true)
-     *                       rhs_info.k0: 2,3,4,8,16 (only 4, 8 and 16 if rhs_info.export_to_cl_image = true)
-     *                       rhs_info.transpose: true
-     * @param[in]  gemm_info GEMM information used to retrieve the original dimensions of the input matrices
-     *
-     * @note lhs_info.k0 must be equal to rhs_info.k0
-     */
-    void configure(const ICLTensor *input0, const ICLTensor *input1, const ICLTensor *input2, ICLTensor *output, float alpha, float beta, const GEMMLHSMatrixInfo &lhs_info,
-                   const GEMMRHSMatrixInfo &rhs_info,
-                   const GEMMKernelInfo    &gemm_info);
-    /** Initialise the kernel's input and output.
-     *
-     * @note The F16 computation also supports mixed precision through the gemm_info.fp_mixed_precision flag.
-     *       Mixed precision combines different floating precisions during the computation, in particular, F32 for the accumulations and F16 for the
-     *       multiplications. i.e. float c = (half)a * (half)b
-     *
-     * @note If rhs_info.export_to_cl_image = true, this OpenCL kernel will fetch the RHS data using the OpenCL read_image built-in function.
-     *       Reading from the OpenCL image object can increase the performance. However, since the OpenCL image object is created importing the OpenCL buffer,
-     *       the following conditions are required:
-     *       -# rhs_info.n0 can only be 4, 8 and 16
-     *       -# rhs_info.k0 can only be 4, 8 and 16
-     *       -# Data type can only be F32
-     *       -# The platform should support the OpenCL cl_khr_image2d_from_buffer extension
-     *       -# The stride Y for the input1 should satisfy the OpenCL pitch alignment requirement
-     *       -# input1 width should be less or equal to (CL_DEVICE_IMAGE2D_MAX_WIDTH * 4)
-     *       -# input1 (height * depth) should be less or equal to CL_DEVICE_IMAGE2D_MAX_HEIGHT
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input0          Input tensor containing the LHS reshaped matrix. Data type supported: F16/F32  (only F32 if rhs_info.export_to_cl_image = true). The number of dimensions for the LHS matrix must be less or equal than 4
-     * @param[in]  input1          Input tensor containing the RHS reshaped matrix. Data type supported: same as @p input0. The number of dimensions for the RHS matrix must be less or equal than 3
-     * @param[in]  input2          Input tensor containing the bias matrix. Data type supported: same as @p input0.
-     * @param[out] output          Output tensor to store the result of matrix multiplication. Data type supported: same as @p input0
-     * @param[in]  alpha           Weight of the matrix product
-     * @param[in]  beta            Weight of the matrix bias
-     * @param[in]  lhs_info        LHS matrix information used for reshaping the input0 tensor.  Only the following values are supported:
-     *                             lhs_info.m0: 2,3,4,5,6,7,8
-     *                             lhs_info.k0: 2,3,4,8,16
-     *                             lhs_info.transpose: false
-     * @param[in]  rhs_info        RHS matrix information used for reshaping the input1 tensor.  Only the following values are supported:
-     *                             rhs_info.n0: 2,3,4,8,16 (only 4, 8 and 16 if rhs_info.export_to_cl_image = true)
-     *                             rhs_info.k0: 2,3,4,8,16 (only 4, 8 and 16 if rhs_info.export_to_cl_image = true)
-     *                             rhs_info.transpose: true
-     * @param[in]  gemm_info       GEMM information used to retrieve the original dimensions of the input matrices
-     *
-     * @note lhs_info.k0 must be equal to rhs_info.k0
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input0, const ICLTensor *input1, const ICLTensor *input2, ICLTensor *output, float alpha, float beta,
-                   const GEMMLHSMatrixInfo &lhs_info,
-                   const GEMMRHSMatrixInfo &rhs_info,
-                   const GEMMKernelInfo    &gemm_info);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLGEMMMatrixMultiplyReshapedKernel
-     *
-     * @note The F16 computation also supports mixed precision through the gemm_info.fp_mixed_precision flag.
-     *       Mixed precision combines different floating precisions during the computation, in particular, F32 for the accumulations and F16 for the
-     *       multiplications. i.e. float c = (half)a * (half)b
-     *
-     * @note If rhs_info.export_to_cl_image = true, this OpenCL kernel will fetch the RHS data using the OpenCL read_image built-in function.
-     *       Reading from the OpenCL image object can increase the performance. However, since the OpenCL image object is created importing the OpenCL buffer,
-     *       the following conditions are required:
-     *       -# rhs_info.n0 can only be 4, 8 and 16
-     *       -# rhs_info.k0 can only be 4, 8 and 16
-     *       -# Data type can only be F32
-     *       -# The platform should support the OpenCL cl_khr_image2d_from_buffer extension
-     *       -# The stride Y for the input1 should satisfy the OpenCL pitch alignment requirement
-     *       -# input1 width should be less or equal to (CL_DEVICE_IMAGE2D_MAX_WIDTH * 4)
-     *       -# input1 (height * depth) should be less or equal to CL_DEVICE_IMAGE2D_MAX_HEIGHT
-     *
-     * @param[in] input0    Input tensor containing the LHS reshaped matrix. Data type supported: F16/F32  (only F32 if rhs_info.export_to_cl_image = true). The number of dimensions for the LHS matrix must be less or equal than 4
-     * @param[in] input1    Input tensor containing the RHS reshaped matrix. Data type supported: same as @p input0. The number of dimensions for the RHS matrix must be less or equal than 3
-     * @param[in] input2    Input tensor info containing the bias matrix. Data type supported: same as @p input0.
-     * @param[in] output    Output tensor to store the result of matrix multiplication. Data type supported: same as @p input0
-     * @param[in] alpha     Weight of the matrix product
-     * @param[in] beta      Weight of the matrix bias
-     * @param[in] lhs_info  LHS matrix information used for reshaping the input0 tensor.  Only the following values are supported:
-     *                      lhs_info.m0: 2,3,4,5,6,7,8
-     *                      lhs_info.k0: 2,3,4,8,16
-     *                      lhs_info.transpose: false
-     * @param[in] rhs_info  RHS matrix information used for reshaping the input1 tensor.  Only the following values are supported:
-     *                      rhs_info.n0: 2,3,4,8,16 (only 4, 8 and 16 if rhs_info.export_to_cl_image = true)
-     *                      rhs_info.k0: 2,3,4,8,16 (only 4, 8 and 16 if rhs_info.export_to_cl_image = true)
-     *                      rhs_info.transpose: true
-     * @param[in] gemm_info GEMM information used to retrieve the original dimensions of the input matrices
-     *
-     * @note lhs_info.k0 must be equal to rhs_info.k0
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input0, const ITensorInfo *input1, const ITensorInfo *input2, const ITensorInfo *output, float alpha, float beta, const GEMMLHSMatrixInfo &lhs_info,
-                           const GEMMRHSMatrixInfo &rhs_info,
-                           const GEMMKernelInfo    &gemm_info);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input0;
-    const ICLTensor *_input1;
-    const ICLTensor *_input2;
-    ICLTensor       *_output;
-    bool             _slide_matrix_b;
-    bool             _reinterpret_output_as_3d;
-    bool             _use_dummy_work_items;
-    bool             _add_bias;
-    bool             _broadcast_bias;
-    bool             _export_to_cl_image;
-    unsigned int     _k;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLGEMMMATRIXMULTIPLYRESHAPEDKERNEL_H*/
\ No newline at end of file
diff --git a/arm_compute/core/CL/kernels/CLGEMMMatrixMultiplyReshapedOnlyRHSKernel.h b/arm_compute/core/CL/kernels/CLGEMMMatrixMultiplyReshapedOnlyRHSKernel.h
deleted file mode 100644
index eab7fd219e..0000000000
--- a/arm_compute/core/CL/kernels/CLGEMMMatrixMultiplyReshapedOnlyRHSKernel.h
+++ /dev/null
@@ -1,168 +0,0 @@
-/*
- * Copyright (c) 2019-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLGEMMMATRIXMULTIPLYRESHAPEDONLYRHSKERNEL_H
-#define ARM_COMPUTE_CLGEMMMATRIXMULTIPLYRESHAPEDONLYRHSKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-#include "arm_compute/core/KernelDescriptors.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** OpenCL kernel to multiply matrices when only the input matrix RHS (input1) has been reshaped
- *
- * @note The input matrix input1 must be reshaped through @ref CLGEMMReshapeRHSMatrixKernel
- */
-class CLGEMMMatrixMultiplyReshapedOnlyRHSKernel : public ICLKernel
-{
-public:
-    /** Default Constructor */
-    CLGEMMMatrixMultiplyReshapedOnlyRHSKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLGEMMMatrixMultiplyReshapedOnlyRHSKernel(const CLGEMMMatrixMultiplyReshapedOnlyRHSKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLGEMMMatrixMultiplyReshapedOnlyRHSKernel &operator=(const CLGEMMMatrixMultiplyReshapedOnlyRHSKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLGEMMMatrixMultiplyReshapedOnlyRHSKernel(CLGEMMMatrixMultiplyReshapedOnlyRHSKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLGEMMMatrixMultiplyReshapedOnlyRHSKernel &operator=(CLGEMMMatrixMultiplyReshapedOnlyRHSKernel &&) = default;
-    /** Initialise the kernel's input and output.
-     *
-     * @note If rhs_info.export_to_cl_image = true, this OpenCL kernel will fetch the RHS data using the OpenCL read_image built-in function.
-     *       Reading from the OpenCL image object can increase the performance. However, since the OpenCL image object is created importing the OpenCL buffer,
-     *       the following conditions are required:
-     *       -# rhs_info.n0 can only be 4, 8 and 16
-     *       -# rhs_info.k0 can only be 4, 8 and 16
-     *       -# Data type can only be F32
-     *       -# The platform should support the OpenCL cl_khr_image2d_from_buffer extension
-     *       -# The stride Y for the input1 should satisfy the OpenCL pitch alignment requirement
-     *       -# input1 width should be less or equal to (CL_DEVICE_IMAGE2D_MAX_WIDTH * 4)
-     *       -# input1 (height * depth) should be less or equal to CL_DEVICE_IMAGE2D_MAX_HEIGHT
-     *
-     * @param[in]  input0    Input tensor containing the LHS matrix. Data type supported: F16/F32 (only F32 if rhs_info.export_to_cl_image = true).
-     *                       The number of dimensions for the LHS matrix must be less or equal than 4.
-     * @param[in]  input1    Input tensor containing the RHS reshaped matrix. Data type supported: same as @p input0. The number of dimensions for the RHS matrix must be less or equal than 3.
-     * @param[in]  input2    Input tensor containing the bias matrix. Data type supported: same as @p input0.
-     * @param[out] output    Output tensor to store the result of matrix multiplication. Data type supported: same as @p input0
-     * @param[in]  alpha     Weight of the matrix product
-     * @param[in]  beta      Weight of the matrix bias
-     * @param[in]  lhs_info  LHS matrix information used to retrieve the number of rows to be processed by each thread. Only the following values are supported:
-     *                       lhs_info.m0: 1,2,3,4,5,6,7,8
-     * @param[in]  rhs_info  RHS matrix information used for reshaping the input1 tensor.  Only the following values are supported:
-     *                       rhs_info.k0: 2,3,4,8,16
-     *                       rhs_info.n0: 2,3,4,8,16
-     *                       rhs_info.transpose: true,false
-     * @param[in]  gemm_info GEMM information used to retrieve the original dimensions of the input matrices
-     */
-    void configure(const ICLTensor *input0, const ICLTensor *input1, const ICLTensor *input2, ICLTensor *output, float alpha, float beta, const GEMMLHSMatrixInfo &lhs_info,
-                   const GEMMRHSMatrixInfo &rhs_info,
-                   const GEMMKernelInfo    &gemm_info);
-    /** Initialise the kernel's input and output.
-     *
-     * @note If rhs_info.export_to_cl_image = true, this OpenCL kernel will fetch the RHS data using the OpenCL read_image built-in function.
-     *       Reading from the OpenCL image object can increase the performance. However, since the OpenCL image object is created importing the OpenCL buffer,
-     *       the following conditions are required:
-     *       -# rhs_info.n0 can only be 4, 8 and 16
-     *       -# rhs_info.k0 can only be 4, 8 and 16
-     *       -# Data type can only be F32
-     *       -# The platform should support the OpenCL cl_khr_image2d_from_buffer extension
-     *       -# The stride Y for the input1 should satisfy the OpenCL pitch alignment requirement
-     *       -# input1 width should be less or equal to (CL_DEVICE_IMAGE2D_MAX_WIDTH * 4)
-     *       -# input1 (height * depth) should be less or equal to CL_DEVICE_IMAGE2D_MAX_HEIGHT
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input0          Input tensor containing the LHS matrix. Data type supported: F16/F32 (only F32 if rhs_info.export_to_cl_image = true).
-     *                             The number of dimensions for the LHS matrix must be less or equal than 4.
-     * @param[in]  input1          Input tensor containing the RHS reshaped matrix. Data type supported: same as @p input0. The number of dimensions for the RHS matrix must be less or equal than 3.
-     * @param[in]  input2          Input tensor containing the bias matrix. Data type supported: same as @p input0.
-     * @param[out] output          Output tensor to store the result of matrix multiplication. Data type supported: same as @p input0
-     * @param[in]  alpha           Weight of the matrix product
-     * @param[in]  beta            Weight of the matrix bias
-     * @param[in]  lhs_info        LHS matrix information used to retrieve the number of rows to be processed by each thread. Only the following values are supported:
-     *                             lhs_info.m0: 1,2,3,4,5,6,7,8
-     * @param[in]  rhs_info        RHS matrix information used for reshaping the input1 tensor.  Only the following values are supported:
-     *                             rhs_info.k0: 2,3,4,8,16
-     *                             rhs_info.n0: 2,3,4,8,16
-     *                             rhs_info.transpose: true,false
-     * @param[in]  gemm_info       GEMM information used to retrieve the original dimensions of the input matrices
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input0, const ICLTensor *input1, const ICLTensor *input2, ICLTensor *output, float alpha, float beta,
-                   const GEMMLHSMatrixInfo &lhs_info,
-                   const GEMMRHSMatrixInfo &rhs_info,
-                   const GEMMKernelInfo    &gemm_info);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLGEMMMatrixMultiplyReshapedOnlyRHSKernel
-     *
-     * @note If rhs_info.export_to_cl_image = true, this OpenCL kernel will fetch the RHS data using the OpenCL read_image built-in function.
-     *       Reading from the OpenCL image object can increase the performance. However, since the OpenCL image object is created importing the OpenCL buffer,
-     *       the following conditions are required:
-     *       -# rhs_info.n0 can only be 4, 8 and 16
-     *       -# rhs_info.k0 can only be 4, 8 and 16
-     *       -# Data type can only be F32
-     *       -# The platform should support the OpenCL cl_khr_image2d_from_buffer extension
-     *       -# The stride Y for the input1 should satisfy the OpenCL pitch alignment requirement
-     *       -# input1 width should be less or equal to (CL_DEVICE_IMAGE2D_MAX_WIDTH * 4)
-     *       -# input1 (height * depth) should be less or equal to CL_DEVICE_IMAGE2D_MAX_HEIGHT
-     *
-     * @param[in] input0    Input tensor info for the LHS matrix. Data type supported: F16/F32 (only F32 if rhs_info.export_to_cl_image = true).
-     *                      The number of dimensions for the LHS matrix must be less or equal than 4.
-     * @param[in] input1    Input tensor info for the RHS reshaped matrix. Data type supported: same as @p input0. The number of dimensions for the RHS matrix must be less or equal than 3.
-     * @param[in] input2    Input tensor info containing the bias matrix. Data type supported: same as @p input0.
-     * @param[in] output    Output tensor info. Data type supported: same as @p input0
-     * @param[in] alpha     Weight of the matrix product
-     * @param[in] beta      Weight of the matrix bias
-     * @param[in] lhs_info  LHS matrix information used to retrieve the number of rows to be processed by each thread. Only the following values are supported:
-     *                      lhs_info.m0: 1,2,3,4,5,6,7,8
-     * @param[in] rhs_info  RHS matrix information used for reshaping the input1 tensor.  Only the following values are supported:
-     *                      rhs_info.k0: 2,3,4,8,16
-     *                      rhs_info.n0: 2,3,4,8,16
-     *                      rhs_info.transpose: true,false
-     * @param[in] gemm_info GEMM information used to retrieve the original dimensions of the input matrices
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input0, const ITensorInfo *input1, const ITensorInfo *input2, const ITensorInfo *output, float alpha, float beta, const GEMMLHSMatrixInfo &lhs_info,
-                           const GEMMRHSMatrixInfo &rhs_info,
-                           const GEMMKernelInfo    &gemm_info);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input0;
-    const ICLTensor *_input1;
-    const ICLTensor *_input2;
-    ICLTensor       *_output;
-    bool             _slide_matrix_b;
-    bool             _reinterpret_input_as_3d;
-    bool             _reinterpret_output_as_3d;
-    bool             _use_dummy_work_items;
-    bool             _add_bias;
-    bool             _broadcast_bias;
-    bool             _export_to_cl_image;
-    bool             _has_pad_y;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLGEMMMATRIXMULTIPLYRESHAPEDONLYRHSKERNEL_H*/
diff --git a/arm_compute/core/CL/kernels/CLGEMMMatrixVectorMultiplyKernel.h b/arm_compute/core/CL/kernels/CLGEMMMatrixVectorMultiplyKernel.h
deleted file mode 100644
index 95ed87d95b..0000000000
--- a/arm_compute/core/CL/kernels/CLGEMMMatrixVectorMultiplyKernel.h
+++ /dev/null
@@ -1,84 +0,0 @@
-/*
- * Copyright (c) 2017-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLGEMMMATRIXVECTORMULTIPLYKERNEL_H
-#define ARM_COMPUTE_CLGEMMMATRIXVECTORMULTIPLYKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the GEMM matrix vector multiply kernel. **/
-class CLGEMMMatrixVectorMultiplyKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLGEMMMatrixVectorMultiplyKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLGEMMMatrixVectorMultiplyKernel(const CLGEMMMatrixVectorMultiplyKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLGEMMMatrixVectorMultiplyKernel &operator=(const CLGEMMMatrixVectorMultiplyKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLGEMMMatrixVectorMultiplyKernel(CLGEMMMatrixVectorMultiplyKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLGEMMMatrixVectorMultiplyKernel &operator=(CLGEMMMatrixVectorMultiplyKernel &&) = default;
-    /** Set the input and output of the kernel.
-     *
-     * @param[in]  input0 The reshaped input tensor. Data types supported: QASYMM8/QASYMM8_SIGNED/F16/F32
-     * @param[in]  input1 The 2D reshaped weights tensor. Data type supported: Same as @p input.
-     * @param[out] output The output 2D tensor. Data types supported: Same as @p input, S32 for QASYMM8/QASYMM8_SIGNED.
-     */
-    void configure(const ICLTensor *input0, const ICLTensor *input1, ICLTensor *output);
-    /** Set the input and output of the kernel.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input0          The reshaped input tensor. Data types supported: QASYMM8/QASYMM8_SIGNED/F16/F32
-     * @param[in]  input1          The 2D reshaped weights tensor. Data type supported: Same as @p input.
-     * @param[out] output          The output 2D tensor. Data types supported: Same as @p input, S32 for QASYMM8/QASYMM8_SIGNED.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input0, const ICLTensor *input1, ICLTensor *output);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLGEMMMatrixVectorMultiplyKernel
-     *
-     * @param[in] input0 The reshaped input tensor info. Data types supported: QASYMM8/QASYMM8_SIGNED/F16/F32
-     * @param[in] input1 The 2D reshaped weights tensor info. Data type supported: Same as @p input.
-     * @param[in] output The output 2D tensor info. Data types supported: Same as @p input, S32 for QASYMM8/QASYMM8_SIGNED.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input0, const ITensorInfo *input1, const ITensorInfo *output);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-    BorderSize border_size() const override;
-
-private:
-    const ICLTensor *_input0;
-    const ICLTensor *_input1;
-    ICLTensor       *_output;
-    int              _num_rows_read_per_iteration;
-    BorderSize       _border_size;
-};
-} // arm_compute
-#endif /*ARM_COMPUTE_CLGEMMMATRIXVECTORMULTIPLYKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLGEMMReshapeLHSMatrixKernel.h b/arm_compute/core/CL/kernels/CLGEMMReshapeLHSMatrixKernel.h
deleted file mode 100644
index 0f74cb85e4..0000000000
--- a/arm_compute/core/CL/kernels/CLGEMMReshapeLHSMatrixKernel.h
+++ /dev/null
@@ -1,105 +0,0 @@
-/*
- * Copyright (c) 2018-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLGEMMRESHAPELHSMATRIXKERNEL_H
-#define ARM_COMPUTE_CLGEMMRESHAPELHSMATRIXKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** OpenCL kernel to reshape the LHS matrix when performing the matrix multiplication.
- *  In particular, this function splits the input matrix in blocks of size M0xK0 (defined through GEMMLHSInfo) and
- *  stores each one in the output matrix unrolling the values
- */
-class CLGEMMReshapeLHSMatrixKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLGEMMReshapeLHSMatrixKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLGEMMReshapeLHSMatrixKernel(const CLGEMMReshapeLHSMatrixKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLGEMMReshapeLHSMatrixKernel &operator=(const CLGEMMReshapeLHSMatrixKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLGEMMReshapeLHSMatrixKernel(CLGEMMReshapeLHSMatrixKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLGEMMReshapeLHSMatrixKernel &operator=(CLGEMMReshapeLHSMatrixKernel &&) = default;
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in]  input                   Input tensor. Data types supported: All
-     * @param[out] output                  Output tensor. Data type supported: same as @p input
-     * @param[in]  lhs_info                LHS matrix information to be used for reshaping. This object contains all the necessary
-     *                                     information to reshape the input tensor. Only the following values are supported:
-     *                                     lhs_info.m0: 2,3,4,5,6,7,8
-     *                                     lhs_info.k0: 2,3,4,8,16
-     *                                     lhs_info.v0: greater than 0
-     *                                     lhs_info.transpose: true, false
-     *                                     lhs_info.interleave: true, false
-     * @param[in]  reinterpret_input_as_3d (Optional) True if the input has to be reinterpreted as 3D tensor
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, const GEMMLHSMatrixInfo &lhs_info, bool reinterpret_input_as_3d = false);
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in]  compile_context         The compile context to be used.
-     * @param[in]  input                   Input tensor. Data types supported: All
-     * @param[out] output                  Output tensor. Data type supported: same as @p input
-     * @param[in]  lhs_info                LHS matrix information to be used for reshaping. This object contains all the necessary
-     *                                     information to reshape the input tensor. Only the following values are supported:
-     *                                     lhs_info.m0: 2,3,4,5,6,7,8
-     *                                     lhs_info.k0: 2,3,4,8,16
-     *                                     lhs_info.v0: greater than 0
-     *                                     lhs_info.transpose: true, false
-     *                                     lhs_info.interleave: true, false
-     * @param[in]  reinterpret_input_as_3d (Optional) True if the input has to be reinterpreted as 3D tensor
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, const GEMMLHSMatrixInfo &lhs_info, bool reinterpret_input_as_3d = false);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLGEMMReshapeLHSMatrixKernel
-     *
-     * @param[in] input                   Input tensor info. Data types supported: All
-     * @param[in] output                  Output tensor info which stores the interleaved matrix. Data type supported: same as @p input.
-     * @param[in] lhs_info                LHS matrix information to be used for reshaping. This object contains all the necessary
-     *                                    information to reshape the input tensor. Only the following values are supported:
-     *                                    lhs_info.m0: 2,3,4,5,6,7,8
-     *                                    lhs_info.k0: 2,3,4,8,16
-     *                                    lhs_info.v0: greater than 0
-     *                                    lhs_info.transpose: true, false
-     *                                    lhs_info.interleave: true, false
-     * @param[in] reinterpret_input_as_3d True if the input has to be reinterpreted as 3D tensor
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output, const GEMMLHSMatrixInfo &lhs_info, bool reinterpret_input_as_3d);
-
-    // Inherited methods overridden
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;
-    ICLTensor       *_output;
-    bool             _reinterpret_input_as_3d;
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLGEMMRESHAPELHSMATRIXKERNEL_H */
\ No newline at end of file
diff --git a/arm_compute/core/CL/kernels/CLGEMMReshapeRHSMatrixKernel.h b/arm_compute/core/CL/kernels/CLGEMMReshapeRHSMatrixKernel.h
deleted file mode 100644
index 557f71b07d..0000000000
--- a/arm_compute/core/CL/kernels/CLGEMMReshapeRHSMatrixKernel.h
+++ /dev/null
@@ -1,133 +0,0 @@
-/*
- * Copyright (c) 2018-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLGEMMRESHAPERHSMATRIXKERNEL_H
-#define ARM_COMPUTE_CLGEMMRESHAPERHSMATRIXKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** OpenCL kernel to reshape the RHS matrix when performing the matrix multiplication
- *  In particular, this kernel splits the input matrix in blocks of size K0xN0 and stores each one in
- *  the output matrix unrolling the values */
-class CLGEMMReshapeRHSMatrixKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLGEMMReshapeRHSMatrixKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLGEMMReshapeRHSMatrixKernel(const CLGEMMReshapeRHSMatrixKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLGEMMReshapeRHSMatrixKernel &operator=(const CLGEMMReshapeRHSMatrixKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLGEMMReshapeRHSMatrixKernel(CLGEMMReshapeRHSMatrixKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLGEMMReshapeRHSMatrixKernel &operator=(CLGEMMReshapeRHSMatrixKernel &&) = default;
-    /** Initialise the kernel's input and output.
-     *
-     * @note If rhs_info.export_to_cl_image = true, this OpenCL kernel will guarantee the OpenCL pitch alignment for the output tensor,
-     *       required to create a OpenCL image object from buffer in @ref CLGEMMMatrixMultiplyReshapedKernel and in @ref CLGEMMMatrixMultiplyReshapedOnlyRHSKernel
-     *       Since the OpenCL image object is created importing the OpenCL buffer, the following conditions are required:
-     *       -# rhs_info.n0 can only be 4, 8 and 16
-     *       -# rhs_info.k0 can only be 4, 8 and 16
-     *       -# Data type can only be F32, F16
-     *       -# The platform should support the OpenCL cl_khr_image2d_from_buffer extension
-     *       -# output width should be less or equal to (CL_DEVICE_IMAGE2D_MAX_WIDTH * 4)
-     *       -# output (height * depth) should be less or equal to CL_DEVICE_IMAGE2D_MAX_HEIGHT
-     *       -# The output tensor should be only consumed by @ref CLGEMMMatrixMultiplyReshapedKernel or @ref CLGEMMMatrixMultiplyReshapedOnlyRHSKernel
-     *
-     * @param[in]  input    Input tensor. Data types supported: All
-     * @param[out] output   Output tensor. Data type supported: same as @p input
-     * @param[in]  rhs_info RHS matrix information to be used for reshaping. This object contains all the necessary
-     *                      information to reshape the input tensor. Only the following values are supported:
-     *                      rhs_info.n0: 2,3,4,8,16 (only 4, 8 and 16 if rhs_info.export_to_cl_image == true)
-     *                      rhs_info.k0: 1,2,3,4,8,16 (k0 = 1 only if rhs_info.transpose = false), (only 4, 8 and 16 if rhs_info.export_to_cl_image == true)
-     *                      rhs_info.h0: greater than 0
-     *                      rhs_info.transpose: true, false
-     *                      rhs_info.interleave: true, false
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, const GEMMRHSMatrixInfo &rhs_info);
-    /** Initialise the kernel's input and output.
-     *
-     * @note If rhs_info.export_to_cl_image = true, this OpenCL kernel will guarantee the OpenCL pitch alignment for the output tensor,
-     *       required to create a OpenCL image object from buffer in @ref CLGEMMMatrixMultiplyReshapedKernel and in @ref CLGEMMMatrixMultiplyReshapedOnlyRHSKernel
-     *       Since the OpenCL image object is created importing the OpenCL buffer, the following conditions are required:
-     *       -# rhs_info.n0 can only be 4, 8 and 16
-     *       -# rhs_info.k0 can only be 4, 8 and 16
-     *       -# Data type can only be F32, F16
-     *       -# The platform should support the OpenCL cl_khr_image2d_from_buffer extension
-     *       -# output width should be less or equal to (CL_DEVICE_IMAGE2D_MAX_WIDTH * 4)
-     *       -# output (height * depth) should be less or equal to CL_DEVICE_IMAGE2D_MAX_HEIGHT
-     *       -# The output tensor should be only consumed by @ref CLGEMMMatrixMultiplyReshapedKernel or @ref CLGEMMMatrixMultiplyReshapedOnlyRHSKernel
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Input tensor. Data types supported: All
-     * @param[out] output          Output tensor. Data type supported: same as @p input
-     * @param[in]  rhs_info        RHS matrix information to be used for reshaping. This object contains all the necessary
-     *                             information to reshape the input tensor. Only the following values are supported:
-     *                             rhs_info.n0: 2,3,4,8,16 (only 4, 8 and 16 if rhs_info.export_to_cl_image == true)
-     *                             rhs_info.k0: 1,2,3,4,8,16 (k0 = 1 only if rhs_info.transpose = false), (only 4, 8 and 16 if rhs_info.export_to_cl_image == true)
-     *                             rhs_info.h0: greater than 0
-     *                             rhs_info.transpose: true, false
-     *                             rhs_info.interleave: true, false
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, const GEMMRHSMatrixInfo &rhs_info);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLGEMMReshapeRHSMatrixKernel
-     *
-     * @note If rhs_info.export_to_cl_image = true, this OpenCL kernel will guarantee the OpenCL pitch alignment for the output tensor,
-     *       required to create a OpenCL image object from buffer in @ref CLGEMMMatrixMultiplyReshapedKernel and in @ref CLGEMMMatrixMultiplyReshapedOnlyRHSKernel
-     *       Since the OpenCL image object is created importing the OpenCL buffer, the following conditions are required:
-     *       -# rhs_info.n0 can only be 4, 8 and 16
-     *       -# rhs_info.k0 can only be 4, 8 and 16
-     *       -# Data type can only be F32, F16
-     *       -# The platform should support the OpenCL cl_khr_image2d_from_buffer extension
-     *       -# output width should be less or equal to (CL_DEVICE_IMAGE2D_MAX_WIDTH * 4)
-     *       -# output (height * depth) should be less or equal to CL_DEVICE_IMAGE2D_MAX_HEIGHT
-     *       -# The output tensor should be only consumed by @ref CLGEMMMatrixMultiplyReshapedKernel or @ref CLGEMMMatrixMultiplyReshapedOnlyRHSKernel
-     *
-     * @param[in] input    Input tensor info. Data types supported: All
-     * @param[in] output   Output tensor info which stores the interleaved matrix. Data type supported: same as @p input.
-     * @param[in] rhs_info RHS matrix information to be used for reshaping. This object contains all the necessary
-     *                     information to reshape the input tensor. Only the following values are supported:
-     *                     rhs_info.n0: 2,3,4,8,16 (only 4, 8 and 16 if rhs_info.export_to_cl_image == true)
-     *                     rhs_info.k0: 1,2,3,4,8,16 (k0 = 1 only if rhs_info.transpose = false),(only 4, 8 and 16 if rhs_info.export_to_cl_image == true)
-     *                     rhs_info.h0: greater than 0
-     *                     rhs_info.transpose: true, false
-     *                     rhs_info.interleave: true, false
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output, const GEMMRHSMatrixInfo &rhs_info);
-
-    // Inherited methods overridden
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;
-    ICLTensor       *_output;
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLGEMMRESHAPERHSMATRIXKERNEL_H */
\ No newline at end of file
diff --git a/arm_compute/core/CL/kernels/CLGatherKernel.h b/arm_compute/core/CL/kernels/CLGatherKernel.h
deleted file mode 100644
index c8a96327b6..0000000000
--- a/arm_compute/core/CL/kernels/CLGatherKernel.h
+++ /dev/null
@@ -1,89 +0,0 @@
-/*
- * Copyright (c) 2018-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLGATHERKERNEL_H
-#define ARM_COMPUTE_CLGATHERKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-#include "arm_compute/core/Types.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the kernel to perform tensor reshaping */
-class CLGatherKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLGatherKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLGatherKernel(const CLGatherKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLGatherKernel &operator=(const CLGatherKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLGatherKernel(CLGatherKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLGatherKernel &operator=(CLGatherKernel &&) = default;
-    /** Default destructor */
-    ~CLGatherKernel() = default;
-    /** Initialise the kernel's inputs and outputs
-     *
-     * @param[in]  input   Source tensor. Supported tensor rank: up to 4. Data type supported: All.
-     * @param[in]  indices Indices tensor. Supported tensor rank: up to 1. Must be one of the following types: U32/S32. Each value must be in range [0, input.shape[@p axis])
-     * @param[out] output  Destination tensor. Data type supported: Same as @p input
-     * @param[in]  axis    (Optional) The axis in @p input to gather @p indices from. Negative values wrap around. Defaults to 0
-     */
-    void configure(const ICLTensor *input, const ICLTensor *indices, ICLTensor *output, int axis = 0);
-    /** Initialise the kernel's inputs and outputs
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Source tensor. Supported tensor rank: up to 4. Data type supported: All.
-     * @param[in]  indices         Indices tensor. Supported tensor rank: up to 1. Must be one of the following types: U32/S32. Each value must be in range [0, input.shape[@p axis])
-     * @param[out] output          Destination tensor. Data type supported: Same as @p input
-     * @param[in]  axis            (Optional) The axis in @p input to gather @p indices from. Negative values wrap around. Defaults to 0
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, const ICLTensor *indices, ICLTensor *output, int axis = 0);
-
-    /** Static function to check if given info will lead to a valid configuration of @ref CLGatherKernel
-     *
-     * @param[in] input   Source tensor info. Supported tensor rank: up to 4. Data type supported: All.
-     * @param[in] indices Indices tensor info. Supported tensor rank: up to 4. Must be one of the following types: U32/S32. Each value must be in range [0, input.shape[@p axis])
-     * @param[in] output  Destination tensor info. Data type supported: Same as @p input
-     * @param[in] axis    (Optional) The axis in @p input to gather @p indices from. Negative values wrap around. Defaults to 0
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *indices, const ITensorInfo *output, int axis = 0);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;   /**< Source tensor */
-    const ICLTensor *_indices; /**< Indices tensor */
-    ICLTensor       *_output;  /**< Destination tensor */
-    int              _axis;    /**< Axis index */
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLGATHERKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLGaussian3x3Kernel.h b/arm_compute/core/CL/kernels/CLGaussian3x3Kernel.h
deleted file mode 100644
index a783527de4..0000000000
--- a/arm_compute/core/CL/kernels/CLGaussian3x3Kernel.h
+++ /dev/null
@@ -1,59 +0,0 @@
-/*
- * Copyright (c) 2016-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLGAUSSIAN3X3KERNEL_H
-#define ARM_COMPUTE_CLGAUSSIAN3X3KERNEL_H
-
-#include "arm_compute/core/CL/ICLSimple2DKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the Gaussian 3x3 filter kernel.
- *
- */
-class CLGaussian3x3Kernel : public ICLSimple2DKernel
-{
-public:
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in]  input            An input tensor. Data types supported: U8
-     * @param[out] output           The output tensor. Data types supported: U8.
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, bool border_undefined);
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in]  compile_context  The compile context to be used.
-     * @param[in]  input            An input tensor. Data types supported: U8
-     * @param[out] output           The output tensor. Data types supported: U8.
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, bool border_undefined);
-
-    // Inherited methods overridden:
-    BorderSize border_size() const override;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLGAUSSIAN3X3KERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLGaussian5x5Kernel.h b/arm_compute/core/CL/kernels/CLGaussian5x5Kernel.h
deleted file mode 100644
index e8c2268e26..0000000000
--- a/arm_compute/core/CL/kernels/CLGaussian5x5Kernel.h
+++ /dev/null
@@ -1,83 +0,0 @@
-/*
- * Copyright (c) 2016-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLGAUSSIAN5X5KERNEL_H
-#define ARM_COMPUTE_CLGAUSSIAN5X5KERNEL_H
-
-#include "arm_compute/core/CL/kernels/CLConvolutionKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the kernel to run the horizontal pass of 5x5 Gaussian filter on a tensor. */
-class CLGaussian5x5HorKernel : public CLSeparableConvolution5x5HorKernel
-{
-public:
-    /** Initialise the kernel's source, destination and border.
-     *
-     * @param[in]  input            Source tensor. Data types supported: U8.
-     * @param[out] output           Destination tensor. Data types supported: S16.
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, bool border_undefined);
-    /** Initialise the kernel's source, destination and border.
-     *
-     * @param[in]  compile_context  The compile context to be used.
-     * @param[in]  input            Source tensor. Data types supported: U8.
-     * @param[out] output           Destination tensor. Data types supported: S16.
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, bool border_undefined);
-
-private:
-    //Make the configure method of the parent class private
-    using CLSeparableConvolution5x5HorKernel::configure;
-};
-
-/** Interface for the kernel to run the vertical pass of 5x5 Gaussian filter on a tensor. */
-class CLGaussian5x5VertKernel : public CLSeparableConvolution5x5VertKernel
-{
-public:
-    /** Initialise the kernel's source, destination and border.
-     *
-     * @param[in]  input            Input tensor(output of horizontal pass). Data types supported: S16.
-     * @param[out] output           Destination tensor. Data types supported: U8.
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, bool border_undefined);
-    /** Initialise the kernel's source, destination and border.
-     *
-     * @param[in]  compile_context  The compile context to be used.
-     * @param[in]  input            Input tensor(output of horizontal pass). Data types supported: S16.
-     * @param[out] output           Destination tensor. Data types supported: U8.
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, bool border_undefined);
-
-private:
-    //Make the configure method of the parent class private
-    using CLSeparableConvolution5x5VertKernel::configure;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLGAUSSIAN5X5KERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLGaussianPyramidKernel.h b/arm_compute/core/CL/kernels/CLGaussianPyramidKernel.h
deleted file mode 100644
index 36e095d4d1..0000000000
--- a/arm_compute/core/CL/kernels/CLGaussianPyramidKernel.h
+++ /dev/null
@@ -1,111 +0,0 @@
-/*
- * Copyright (c) 2017-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLGAUSSIANPYRAMIDKERNEL_H
-#define ARM_COMPUTE_CLGAUSSIANPYRAMIDKERNEL_H
-
-#include "arm_compute/core/CL/ICLSimpleKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** OpenCL kernel to perform a Gaussian filter and half scaling across width (horizontal pass) */
-class CLGaussianPyramidHorKernel : public ICLSimpleKernel
-{
-public:
-    /** Default constructor */
-    CLGaussianPyramidHorKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLGaussianPyramidHorKernel(const CLGaussianPyramidHorKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLGaussianPyramidHorKernel &operator=(const CLGaussianPyramidHorKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLGaussianPyramidHorKernel(CLGaussianPyramidHorKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLGaussianPyramidHorKernel &operator=(CLGaussianPyramidHorKernel &&) = default;
-    /** Default destructor */
-    ~CLGaussianPyramidHorKernel() = default;
-
-    /** Initialise the kernel's source, destination and border mode.
-     *
-     * @param[in]  input  Source tensor. Data types supported: U8.
-     * @param[out] output Destination tensor. Output should have half the input width. Data types supported: U16.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output);
-    /** Initialise the kernel's source, destination and border mode.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Source tensor. Data types supported: U8.
-     * @param[out] output          Destination tensor. Output should have half the input width. Data types supported: U16.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-    BorderSize border_size() const override;
-
-private:
-    int _l2_load_offset;
-};
-
-/** OpenCL kernel to perform a Gaussian filter and half scaling across height (vertical pass) */
-class CLGaussianPyramidVertKernel : public ICLSimpleKernel
-{
-public:
-    /** Default constructor */
-    CLGaussianPyramidVertKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLGaussianPyramidVertKernel(const CLGaussianPyramidVertKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLGaussianPyramidVertKernel &operator=(const CLGaussianPyramidVertKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLGaussianPyramidVertKernel(CLGaussianPyramidVertKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLGaussianPyramidVertKernel &operator=(CLGaussianPyramidVertKernel &&) = default;
-    /** Default destructor */
-    ~CLGaussianPyramidVertKernel() = default;
-
-    /** Initialise the kernel's source, destination and border mode.
-     *
-     * @param[in]  input  Source tensor. Data types supported: U16.
-     * @param[out] output Destination tensor. Output should have half the input height. Data types supported: U8.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output);
-    /** Initialise the kernel's source, destination and border mode.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Source tensor. Data types supported: U16.
-     * @param[out] output          Destination tensor. Output should have half the input height. Data types supported: U8.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-    BorderSize border_size() const override;
-
-private:
-    int _t2_load_offset;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLGAUSSIANPYRAMIDKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLGenerateProposalsLayerKernel.h b/arm_compute/core/CL/kernels/CLGenerateProposalsLayerKernel.h
deleted file mode 100644
index 9dfe4a42ce..0000000000
--- a/arm_compute/core/CL/kernels/CLGenerateProposalsLayerKernel.h
+++ /dev/null
@@ -1,85 +0,0 @@
-/*
- * Copyright (c) 2019-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLGENERATEPROPOSALSLAYERKERNEL_H
-#define ARM_COMPUTE_CLGENERATEPROPOSALSLAYERKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for Compute All Anchors kernel */
-class CLComputeAllAnchorsKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLComputeAllAnchorsKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLComputeAllAnchorsKernel(const CLComputeAllAnchorsKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLComputeAllAnchorsKernel &operator=(const CLComputeAllAnchorsKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLComputeAllAnchorsKernel(CLComputeAllAnchorsKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLComputeAllAnchorsKernel &operator=(CLComputeAllAnchorsKernel &&) = default;
-    /** Default destructor */
-    ~CLComputeAllAnchorsKernel() = default;
-
-    /** Set the input and output tensors.
-     *
-     * @param[in]  anchors     Source tensor. Original set of anchors of size (4, A), where A is the number of anchors. Data types supported: QSYMM16/F16/F32
-     * @param[out] all_anchors Destination tensor. Destination anchors of size (4, H*W*A) where H and W are the height and width of the feature map and A is the number of anchors. Data types supported: Same as @p input
-     * @param[in]  info        Contains Compute Anchors operation information described in @ref ComputeAnchorsInfo
-     *
-     */
-    void configure(const ICLTensor *anchors, ICLTensor *all_anchors, const ComputeAnchorsInfo &info);
-    /** Set the input and output tensors.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  anchors         Source tensor. Original set of anchors of size (4, A), where A is the number of anchors. Data types supported: QSYMM16/F16/F32
-     * @param[out] all_anchors     Destination tensor. Destination anchors of size (4, H*W*A) where H and W are the height and width of the feature map and A is the number of anchors. Data types supported: Same as @p input
-     * @param[in]  info            Contains Compute Anchors operation information described in @ref ComputeAnchorsInfo
-     *
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *anchors, ICLTensor *all_anchors, const ComputeAnchorsInfo &info);
-
-    /** Static function to check if given info will lead to a valid configuration of @ref CLComputeAllAnchorsKernel
-     *
-     * @param[in] anchors     Source tensor info. Original set of anchors of size (4, A), where A is the number of anchors. Data types supported: QSYMM16/F16/F32
-     * @param[in] all_anchors Destination tensor info. Destination anchors of size (4, H*W*A) where H and W are the height and width of the feature map and A is the number of anchors. Data types supported: Same as @p input
-     * @param[in] info        Contains Compute Anchors operation information described in @ref ComputeAnchorsInfo
-     *
-     * @return a Status
-     */
-    static Status validate(const ITensorInfo *anchors, const ITensorInfo *all_anchors, const ComputeAnchorsInfo &info);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_anchors;
-    ICLTensor       *_all_anchors;
-};
-} // arm_compute
-#endif // ARM_COMPUTE_CLGENERATEPROSPOSALSLAYERKERNEL_H
diff --git a/arm_compute/core/CL/kernels/CLHOGDescriptorKernel.h b/arm_compute/core/CL/kernels/CLHOGDescriptorKernel.h
deleted file mode 100644
index c001aa2c9f..0000000000
--- a/arm_compute/core/CL/kernels/CLHOGDescriptorKernel.h
+++ /dev/null
@@ -1,122 +0,0 @@
-/*
- * Copyright (c) 2017-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLHOGDESCRIPTORKERNEL_H
-#define ARM_COMPUTE_CLHOGDESCRIPTORKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-#include "arm_compute/core/IHOG.h"
-#include "arm_compute/core/Size2D.h"
-
-namespace arm_compute
-{
-class ITensor;
-
-/** OpenCL kernel to perform HOG Orientation Binning */
-class CLHOGOrientationBinningKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLHOGOrientationBinningKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLHOGOrientationBinningKernel(const CLHOGOrientationBinningKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLHOGOrientationBinningKernel &operator=(const CLHOGOrientationBinningKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLHOGOrientationBinningKernel(CLHOGOrientationBinningKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLHOGOrientationBinningKernel &operator=(CLHOGOrientationBinningKernel &&) = default;
-    /** Default destructor */
-    ~CLHOGOrientationBinningKernel() = default;
-
-    /**  Initialise the kernel's inputs, output and HOG's metadata
-     *
-     * @param[in]  input_magnitude Input tensor which stores the magnitude of the gradient for each pixel. Data type supported: S16.
-     * @param[in]  input_phase     Input tensor which stores the phase of the gradient for each pixel. Data type supported: U8
-     * @param[out] output          Output tensor which stores the local HOG for each cell. DataType supported: F32. Number of channels supported: equal to the number of histogram bins per cell
-     * @param[in]  hog_info        HOG's metadata
-     */
-    void configure(const ICLTensor *input_magnitude, const ICLTensor *input_phase, ICLTensor *output, const HOGInfo *hog_info);
-    /**  Initialise the kernel's inputs, output and HOG's metadata
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input_magnitude Input tensor which stores the magnitude of the gradient for each pixel. Data type supported: S16.
-     * @param[in]  input_phase     Input tensor which stores the phase of the gradient for each pixel. Data type supported: U8
-     * @param[out] output          Output tensor which stores the local HOG for each cell. DataType supported: F32. Number of channels supported: equal to the number of histogram bins per cell
-     * @param[in]  hog_info        HOG's metadata
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input_magnitude, const ICLTensor *input_phase, ICLTensor *output, const HOGInfo *hog_info);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input_magnitude;
-    const ICLTensor *_input_phase;
-    ICLTensor       *_output;
-    Size2D           _cell_size;
-};
-
-/** OpenCL kernel to perform HOG block normalization */
-class CLHOGBlockNormalizationKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLHOGBlockNormalizationKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLHOGBlockNormalizationKernel(const CLHOGBlockNormalizationKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLHOGBlockNormalizationKernel &operator=(const CLHOGBlockNormalizationKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLHOGBlockNormalizationKernel(CLHOGBlockNormalizationKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLHOGBlockNormalizationKernel &operator=(CLHOGBlockNormalizationKernel &&) = default;
-    /** Default destructor */
-    ~CLHOGBlockNormalizationKernel() = default;
-
-    /** Initialise the kernel's input, output and HOG's metadata
-     *
-     * @param[in]  input    Input tensor which stores the local HOG for each cell. Data type supported: F32. Number of channels supported: equal to the number of histogram bins per cell
-     * @param[out] output   Output tensor which stores the normalised blocks. Data type supported: F32. Number of channels supported: equal to the number of histogram bins per block
-     * @param[in]  hog_info HOG's metadata
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, const HOGInfo *hog_info);
-    /** Initialise the kernel's input, output and HOG's metadata
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Input tensor which stores the local HOG for each cell. Data type supported: F32. Number of channels supported: equal to the number of histogram bins per cell
-     * @param[out] output          Output tensor which stores the normalised blocks. Data type supported: F32. Number of channels supported: equal to the number of histogram bins per block
-     * @param[in]  hog_info        HOG's metadata
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, const HOGInfo *hog_info);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;
-    ICLTensor       *_output;
-    Size2D           _num_cells_per_block_stride;
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLHOGDESCRIPTORKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLHOGDetectorKernel.h b/arm_compute/core/CL/kernels/CLHOGDetectorKernel.h
deleted file mode 100644
index dc9bba8f20..0000000000
--- a/arm_compute/core/CL/kernels/CLHOGDetectorKernel.h
+++ /dev/null
@@ -1,96 +0,0 @@
-/*
- * Copyright (c) 2017-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLHOGDETECTORKERNEL_H
-#define ARM_COMPUTE_CLHOGDETECTORKERNEL_H
-
-#include "arm_compute/core/CL/ICLArray.h"
-#include "arm_compute/core/CL/ICLHOG.h"
-#include "arm_compute/core/CL/ICLKernel.h"
-#include "arm_compute/core/CL/OpenCL.h"
-
-namespace cl
-{
-class Buffer;
-}
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** OpenCL kernel to perform HOG detector kernel using linear SVM */
-class CLHOGDetectorKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLHOGDetectorKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLHOGDetectorKernel(const CLHOGDetectorKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLHOGDetectorKernel &operator=(const CLHOGDetectorKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLHOGDetectorKernel(CLHOGDetectorKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLHOGDetectorKernel &operator=(CLHOGDetectorKernel &&) = default;
-    /** Default destructor */
-    ~CLHOGDetectorKernel() = default;
-
-    /** Initialise the kernel's input, HOG data-object, detection window, the stride of the detection window, the threshold and index of the object to detect
-     *
-     * @param[in]  input                   Input tensor which stores the HOG descriptor obtained with @ref CLHOGOrientationBinningKernel. Data type supported: F32. Number of channels supported: equal to the number of histogram bins per block
-     * @param[in]  hog                     HOG data object used by @ref CLHOGOrientationBinningKernel and  @ref CLHOGBlockNormalizationKernel
-     * @param[out] detection_windows       Array of @ref DetectionWindow. This array stores all the detected objects
-     * @param[in]  num_detection_windows   Number of detected objects
-     * @param[in]  detection_window_stride Distance in pixels between 2 consecutive detection windows in x and y directions.
-     *                                     It must be multiple of the hog->info()->block_stride()
-     * @param[in]  threshold               (Optional) Threshold for the distance between features and SVM classifying plane
-     * @param[in]  idx_class               (Optional) Index of the class used for evaluating which class the detection window belongs to
-     */
-    void configure(const ICLTensor *input, const ICLHOG *hog, ICLDetectionWindowArray *detection_windows, cl::Buffer *num_detection_windows, const Size2D &detection_window_stride, float threshold = 0.0f,
-                   uint16_t idx_class = 0);
-    /** Initialise the kernel's input, HOG data-object, detection window, the stride of the detection window, the threshold and index of the object to detect
-     *
-     * @param[in]  compile_context         The compile context to be used.
-     * @param[in]  input                   Input tensor which stores the HOG descriptor obtained with @ref CLHOGOrientationBinningKernel. Data type supported: F32. Number of channels supported: equal to the number of histogram bins per block
-     * @param[in]  hog                     HOG data object used by @ref CLHOGOrientationBinningKernel and  @ref CLHOGBlockNormalizationKernel
-     * @param[out] detection_windows       Array of @ref DetectionWindow. This array stores all the detected objects
-     * @param[in]  num_detection_windows   Number of detected objects
-     * @param[in]  detection_window_stride Distance in pixels between 2 consecutive detection windows in x and y directions.
-     *                                     It must be multiple of the hog->info()->block_stride()
-     * @param[in]  threshold               (Optional) Threshold for the distance between features and SVM classifying plane
-     * @param[in]  idx_class               (Optional) Index of the class used for evaluating which class the detection window belongs to
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, const ICLHOG *hog, ICLDetectionWindowArray *detection_windows, cl::Buffer *num_detection_windows,
-                   const Size2D &detection_window_stride, float threshold = 0.0f,
-                   uint16_t idx_class = 0);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue);
-
-private:
-    const ICLTensor         *_input;
-    ICLDetectionWindowArray *_detection_windows;
-    cl::Buffer              *_num_detection_windows;
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLHOGDETECTORKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLHarrisCornersKernel.h b/arm_compute/core/CL/kernels/CLHarrisCornersKernel.h
deleted file mode 100644
index 38a2f04adf..0000000000
--- a/arm_compute/core/CL/kernels/CLHarrisCornersKernel.h
+++ /dev/null
@@ -1,100 +0,0 @@
-/*
- * Copyright (c) 2016-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLHARRISCORNERSKERNEL_H
-#define ARM_COMPUTE_CLHARRISCORNERSKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-#include <cstdint>
-
-namespace arm_compute
-{
-class ICLTensor;
-using ICLImage = ICLTensor;
-
-/** Interface for the harris score kernel.
- *
- * @note The implementation supports 3, 5, and 7 for the block_size.
- */
-class CLHarrisScoreKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLHarrisScoreKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLHarrisScoreKernel(const CLHarrisScoreKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLHarrisScoreKernel &operator=(const CLHarrisScoreKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLHarrisScoreKernel(CLHarrisScoreKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLHarrisScoreKernel &operator=(CLHarrisScoreKernel &&) = default;
-    /** Default destructor */
-    ~CLHarrisScoreKernel() = default;
-
-    /** Setup the kernel parameters
-     *
-     * @param[in]  input1           Source image (gradient X). Data types supported S16, S32. (Must be the same as input2)
-     * @param[in]  input2           Source image (gradient Y). Data types supported S16, S32. (Must be the same as input1)
-     * @param[out] output           Destination image (harris score). Data types supported F32
-     * @param[in]  block_size       The block window size used to compute the Harris Corner score.  Supports: 3, 5 and 7
-     * @param[in]  norm_factor      Normalization factor to use accordingly with the gradient size (Must be different from 0)
-     * @param[in]  strength_thresh  Minimum threshold with which to eliminate Harris Corner scores (computed using the normalized Sobel kernel).
-     * @param[in]  sensitivity      Sensitivity threshold k from the Harris-Stephens equation.
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const ICLImage *input1, const ICLImage *input2, ICLImage *output,
-                   int32_t block_size, float norm_factor, float strength_thresh, float sensitivity,
-                   bool border_undefined);
-    /** Setup the kernel parameters
-     *
-     * @param[in]  compile_context  The compile context to be used.
-     * @param[in]  input1           Source image (gradient X). Data types supported S16, S32. (Must be the same as input2)
-     * @param[in]  input2           Source image (gradient Y). Data types supported S16, S32. (Must be the same as input1)
-     * @param[out] output           Destination image (harris score). Data types supported F32
-     * @param[in]  block_size       The block window size used to compute the Harris Corner score.  Supports: 3, 5 and 7
-     * @param[in]  norm_factor      Normalization factor to use accordingly with the gradient size (Must be different from 0)
-     * @param[in]  strength_thresh  Minimum threshold with which to eliminate Harris Corner scores (computed using the normalized Sobel kernel).
-     * @param[in]  sensitivity      Sensitivity threshold k from the Harris-Stephens equation.
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLImage *input1, const ICLImage *input2, ICLImage *output,
-                   int32_t block_size, float norm_factor, float strength_thresh, float sensitivity,
-                   bool border_undefined);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-    BorderSize border_size() const override;
-
-protected:
-    const ICLImage *_input1;          /**< Source image - Gx component */
-    const ICLImage *_input2;          /**< Source image - Gy component */
-    ICLImage       *_output;          /**< Source image - Harris score */
-    float           _sensitivity;     /**< Sensitivity value */
-    float           _strength_thresh; /**< Threshold value */
-    float           _norm_factor;     /**< Normalization factor */
-    BorderSize      _border_size;     /**< Border size */
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLHARRISCORNERSKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLHeightConcatenateLayerKernel.h b/arm_compute/core/CL/kernels/CLHeightConcatenateLayerKernel.h
deleted file mode 100644
index f362441944..0000000000
--- a/arm_compute/core/CL/kernels/CLHeightConcatenateLayerKernel.h
+++ /dev/null
@@ -1,77 +0,0 @@
-/*
- * Copyright (c) 2019-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-
-#ifndef ARM_COMPUTE_CLHEIGHTCONCATENATELAYERKERNEL_H
-#define ARM_COMPUTE_CLHEIGHTCONCATENATELAYERKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-#include "arm_compute/core/Types.h"
-
-namespace arm_compute
-{
-/** Interface for the height concatenate kernel.
- *  The input tensor will be concatenated into the output tensor.
- */
-class CLHeightConcatenateLayerKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLHeightConcatenateLayerKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLHeightConcatenateLayerKernel(const CLHeightConcatenateLayerKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLHeightConcatenateLayerKernel &operator=(const CLHeightConcatenateLayerKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLHeightConcatenateLayerKernel(CLHeightConcatenateLayerKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLHeightConcatenateLayerKernel &operator=(CLHeightConcatenateLayerKernel &&) = default;
-    /** Default destructor */
-    ~CLHeightConcatenateLayerKernel() = default;
-    /** Initialise the kernel's inputs and output
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Input tensor. Data types supported: All.
-     * @param[in]  height_offset   The starting offset on the Y axis for the output tensor.
-     * @param[out] output          Output tensor. Data types supported: Same as @p input.
-     *
-     */
-    void configure(const CLCompileContext &compile_context, ITensorInfo *input, unsigned int height_offset, ITensorInfo *output);
-    /**  Static function to check if given info will lead to a valid configuration of @ref CLHeightConcatenateLayerKernel
-     *
-     * @param[in] input         Input tensor info. Data types supported: All.
-     * @param[in] height_offset The starting offset on the Y axis for the output tensor.
-     * @param[in] output        Output tensor info. Data types supported: Same as @p input.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, unsigned int height_offset, const ITensorInfo *output);
-
-    // Inherited methods overridden:
-    void run_op(ITensorPack &tensors, const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    unsigned int _height_offset;
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLHEIGHTCONCATENATELAYERKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLHistogramKernel.h b/arm_compute/core/CL/kernels/CLHistogramKernel.h
deleted file mode 100644
index 7cb79db6e9..0000000000
--- a/arm_compute/core/CL/kernels/CLHistogramKernel.h
+++ /dev/null
@@ -1,111 +0,0 @@
-/*
- * Copyright (c) 2016-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLHISTOGRAMKERNEL_H
-#define ARM_COMPUTE_CLHISTOGRAMKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLDistribution1D;
-class ICLTensor;
-using ICLImage = ICLTensor;
-
-/** Interface to run the histogram kernel. This kernel processes the part of image with width can be divided by 16.
- *  If the image width is not a multiple of 16, remaining pixels have to be processed with the @ref CLHistogramBorderKernel
- */
-class CLHistogramKernel : public ICLKernel
-{
-public:
-    /** Constructor */
-    CLHistogramKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLHistogramKernel(const CLHistogramKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLHistogramKernel &operator=(const CLHistogramKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLHistogramKernel(CLHistogramKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLHistogramKernel &operator=(CLHistogramKernel &&) = default;
-    /** Initialise the kernel's input, output and border mode.
-     *
-     * @param[in]  input  Source image. Data types supported: U8.
-     * @param[out] output Destination distribution.
-     */
-    void configure(const ICLImage *input, ICLDistribution1D *output);
-    /** Initialise the kernel's input, output and border mode.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Source image. Data types supported: U8.
-     * @param[out] output          Destination distribution.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLImage *input, ICLDistribution1D *output);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLImage    *_input;
-    ICLDistribution1D *_output;
-};
-
-/** Interface to run the histogram kernel to handle the leftover part of image
- *
- */
-class CLHistogramBorderKernel : public ICLKernel
-{
-public:
-    /** Constructor */
-    CLHistogramBorderKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLHistogramBorderKernel(const CLHistogramBorderKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLHistogramBorderKernel &operator=(const CLHistogramBorderKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLHistogramBorderKernel(CLHistogramBorderKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLHistogramBorderKernel &operator=(CLHistogramBorderKernel &&) = default;
-    /** Initialise the kernel's input, output and border mode.
-     *
-     * @param[in]  input  Source image. Data types supported: U8.
-     * @param[out] output Destination distribution.
-     */
-    void configure(const ICLImage *input, ICLDistribution1D *output);
-    /** Initialise the kernel's input, output and border mode.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Source image. Data types supported: U8.
-     * @param[out] output          Destination distribution.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLImage *input, ICLDistribution1D *output);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLImage    *_input;
-    ICLDistribution1D *_output;
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLHISTOGRAMKERNEL_H*/
diff --git a/arm_compute/core/CL/kernels/CLIm2ColKernel.h b/arm_compute/core/CL/kernels/CLIm2ColKernel.h
deleted file mode 100644
index 7b7bd03108..0000000000
--- a/arm_compute/core/CL/kernels/CLIm2ColKernel.h
+++ /dev/null
@@ -1,136 +0,0 @@
-/*
- * Copyright (c) 2017-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLIM2COLKERNEL_H
-#define ARM_COMPUTE_CLIM2COLKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-#include "arm_compute/core/Size2D.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the im2col reshape kernel.
- *
- * Rearranges image blocks into columns. It is used to strip out each convolution block to a single column.
- * It is used to transform a convolution to a plain matrix multiplication.
- *
- * For example taking into account the image below and assuming 3x3 image blocks with stride of 1 we have:
- * @f[
- * \left( \begin{array}{cccc}
- * a00 & a01 & a02 & a03 \\
- * a10 & a11 & a12 & a13 \\
- * a20 & a21 & a22 & a23 \\
- * a30 & a31 & a32 & a33 \\
- * \end{array} \right)
- * =
- * \left( \begin{array}{ccccccccc}
- * a00 & a01 & a02 & a10 & a11 & a12 & a20 & a21 & a22 \\
- * a01 & a02 & a03 & a11 & a12 & a13 & a21 & a22 & a23 \\
- * a10 & a11 & a12 & a20 & a21 & a22 & a30 & a31 & a32 \\
- * a11 & a12 & a13 & a21 & a22 & a23 & a31 & a32 & a33 \\
- * \end{array} \right)
- * @f]
- */
-class CLIm2ColKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLIm2ColKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLIm2ColKernel(const CLIm2ColKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLIm2ColKernel &operator=(const CLIm2ColKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLIm2ColKernel(CLIm2ColKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLIm2ColKernel &operator=(CLIm2ColKernel &&) = default;
-    /** Set the input and output of the kernel.
-     *
-     * @param[in]  input       The input tensor to convert. 3 lower dimensions represent a single input [width, height, IFM],
-     *                         while every optional dimension from 4 and above represent a batch of inputs. Data types supported: QASYMM8/QASYMM8_SIGNED/F16/F32
-     * @param[out] output      The output tensor. First 2 lower dimensions represent a transform of each 3D input,
-     *                         while every dimension above represents a batch. Data types supported: Same as @p input
-     * @param[in]  kernel_dims The kernel dimensions (width and height).
-     * @param[in]  conv_info   Contains padding and stride information described in @ref PadStrideInfo.
-     * @param[in]  has_bias    In case biases are provided expands the matrix with 1.
-     *                         This is valid only for non-quantized inputs.
-     * @param[in]  dilation    (Optional) Dilation, in elements, across x and y. Defaults to (1, 1).
-     * @param[in]  num_groups  (Optional) Number of groups when performing a grouped convolution.
-     *                         Number of groups other than 1 is only supported for NCHW data layout.
-     *                         Number of groups should be multiple to the number of channels.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, const Size2D &kernel_dims, const PadStrideInfo &conv_info, bool has_bias, const Size2D &dilation = Size2D(1U, 1U),
-                   unsigned int num_groups = 1);
-    /** Set the input and output of the kernel.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           The input tensor to convert. 3 lower dimensions represent a single input [width, height, IFM],
-     *                             while every optional dimension from 4 and above represent a batch of inputs. Data types supported: QASYMM8/QASYMM8_SIGNED/F16/F32
-     * @param[out] output          The output tensor. First 2 lower dimensions represent a transform of each 3D input,
-     *                             while every dimension above represents a batch. Data types supported: Same as @p input
-     * @param[in]  kernel_dims     The kernel dimensions (width and height).
-     * @param[in]  conv_info       Contains padding and stride information described in @ref PadStrideInfo.
-     * @param[in]  has_bias        In case biases are provided expands the matrix with 1.
-     * @param[in]  dilation        (Optional) Dilation, in elements, across x and y. Defaults to (1, 1).
-     * @param[in]  num_groups      (Optional) Number of groups when performing a grouped convolution. num_groups != 1 is only supported for NCHW data layout
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, const Size2D &kernel_dims, const PadStrideInfo &conv_info, bool has_bias,
-                   const Size2D &dilation   = Size2D(1U, 1U),
-                   unsigned int  num_groups = 1);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLIm2ColKernel
-     *
-     * @param[in] input       The input tensor to convert. 3 lower dimensions represent a single input [width, height, IFM],
-     *                        while every optional dimension from 4 and above represent a batch of inputs. Data types supported: QASYMM8/QASYMM8_SIGNED/F16/F32
-     * @param[in] output      The output tensor. First 2 lower dimensions represent a transform of each 3D input,
-     *                        while every dimension above represents a batch. Data types supported: Same as @p input
-     * @param[in] kernel_dims The kernel dimensions (width and height).
-     * @param[in] conv_info   Contains padding and stride information described in @ref PadStrideInfo.
-     * @param[in] has_bias    In case biases are provided expands the matrix with 1.
-     *                        This is valid only for non-quantized inputs.
-     * @param[in] dilation    (Optional) Dilation, in elements, across x and y. Defaults to (1, 1).
-     * @param[in] num_groups  (Optional) Number of groups when performing a grouped convolution.
-     *                        Number of groups other than 1 is only supported for NCHW data layout.
-     *                        Number of groups should be multiple to the number of channels.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output, const Size2D &kernel_dims, const PadStrideInfo &conv_info, bool has_bias, const Size2D &dilation = Size2D(1U, 1U),
-                           unsigned int num_groups = 1);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-public:
-    const ICLTensor *_input;
-    ICLTensor       *_output;
-    DataLayout       _data_layout;
-    std::pair<unsigned int, unsigned int> _convolved_dims;
-    unsigned int  _num_elems_processed_per_iteration;
-    Size2D        _kernel_dims;
-    PadStrideInfo _conv_info;
-    unsigned int  _num_groups;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLIM2COLKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLInstanceNormalizationLayerKernel.h b/arm_compute/core/CL/kernels/CLInstanceNormalizationLayerKernel.h
deleted file mode 100644
index a3fdd3c4e7..0000000000
--- a/arm_compute/core/CL/kernels/CLInstanceNormalizationLayerKernel.h
+++ /dev/null
@@ -1,90 +0,0 @@
-/*
- * Copyright (c) 2019-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLINSTANCENORMALIZATIONLAYERKERNEL_H
-#define ARM_COMPUTE_CLINSTANCENORMALIZATIONLAYERKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-#include "arm_compute/core/KernelDescriptors.h"
-
-namespace arm_compute
-{
-// Forward declarations
-class ICLTensor;
-
-/** Interface for performing an instance normalization */
-class CLInstanceNormalizationLayerKernel : public ICLKernel
-{
-public:
-    /** Constructor */
-    CLInstanceNormalizationLayerKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLInstanceNormalizationLayerKernel(const CLInstanceNormalizationLayerKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLInstanceNormalizationLayerKernel &operator=(const CLInstanceNormalizationLayerKernel &) = delete;
-    /** Default Move Constructor. */
-    CLInstanceNormalizationLayerKernel(CLInstanceNormalizationLayerKernel &&) = default;
-    /** Default move assignment operator */
-    CLInstanceNormalizationLayerKernel &operator=(CLInstanceNormalizationLayerKernel &&) = default;
-    /** Default destructor */
-    ~CLInstanceNormalizationLayerKernel() = default;
-
-    /** Set the input and output tensors.
-     *
-     * @param[in, out] input  Source tensor. Data types supported: F16/F32. Data layout supported: NCHW, NHWC
-     *                        In case of @p output tensor = nullptr this tensor will store the result of the normalization.
-     * @param[out]     output Destination tensor. Data types and data layouts supported: same as @p input.
-     * @param[in]      info   Kernel meta-data descriptor
-     */
-    void configure(ICLTensor *input, ICLTensor *output, const InstanceNormalizationLayerKernelInfo &info);
-    /** Set the input and output tensors.
-     *
-     * @param[in]      compile_context The compile context to be used.
-     * @param[in, out] input           Source tensor. Data types supported: F16/F32. Data layout supported: NCHW, NHWC
-     *                                 In case of @p output tensor = nullptr this tensor will store the result of the normalization.
-     * @param[out]     output          Destination tensor. Data types and data layouts supported: same as @p input.
-     * @param[in]      info            Kernel meta-data descriptor
-     */
-    void configure(const CLCompileContext &compile_context, ICLTensor *input, ICLTensor *output, const InstanceNormalizationLayerKernelInfo &info);
-
-    /** Static function to check if given info will lead to a valid configuration of @ref CLInstanceNormalizationLayer.
-     *
-     * @param[in] input  Source tensor info. Data types supported: F16/F32. Data layout supported: NHWC, NCHW
-     * @param[in] output Destination tensor info. Data types and data layouts supported: same as @p input.
-     * @param[in] info   Kernel meta-data descriptor
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output, const InstanceNormalizationLayerKernelInfo &info);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    ICLTensor *_input;
-    ICLTensor *_output;
-    bool       _run_in_place;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLINSTANCENORMALIZATIONLAYERKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLIntegralImageKernel.h b/arm_compute/core/CL/kernels/CLIntegralImageKernel.h
deleted file mode 100644
index cef699ab54..0000000000
--- a/arm_compute/core/CL/kernels/CLIntegralImageKernel.h
+++ /dev/null
@@ -1,86 +0,0 @@
-/*
- * Copyright (c) 2016-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLINTEGRALIMAGEKERNEL_H
-#define ARM_COMPUTE_CLINTEGRALIMAGEKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-#include "arm_compute/core/CL/ICLSimple2DKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface to run the horizontal pass of the integral image kernel. */
-class CLIntegralImageHorKernel : public ICLSimple2DKernel
-{
-public:
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in]  input  An input tensor. Data types supported: U8
-     * @param[out] output Destination tensor, Data types supported: U32.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output);
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           An input tensor. Data types supported: U8
-     * @param[out] output          Destination tensor, Data types supported: U32.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output);
-};
-
-/** Interface to run the vertical pass of the integral image kernel. */
-class CLIntegralImageVertKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLIntegralImageVertKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLIntegralImageVertKernel(const CLIntegralImageVertKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLIntegralImageVertKernel &operator=(const CLIntegralImageVertKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLIntegralImageVertKernel(CLIntegralImageVertKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLIntegralImageVertKernel &operator=(CLIntegralImageVertKernel &&) = default;
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in,out] in_out The input/output tensor. Data types supported: U32
-     */
-    void configure(ICLTensor *in_out);
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in]     compile_context The compile context to be used.
-     * @param[in,out] in_out          The input/output tensor. Data types supported: U32
-     */
-    void configure(const CLCompileContext &compile_context, ICLTensor *in_out);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    ICLTensor *_in_out;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLINTEGRALIMAGEKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLL2NormalizeLayerKernel.h b/arm_compute/core/CL/kernels/CLL2NormalizeLayerKernel.h
deleted file mode 100644
index 55fe563954..0000000000
--- a/arm_compute/core/CL/kernels/CLL2NormalizeLayerKernel.h
+++ /dev/null
@@ -1,100 +0,0 @@
-/*
- * Copyright (c) 2017-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLL2NORMALIZELAYERKERNEL_H
-#define ARM_COMPUTE_CLL2NORMALIZELAYERKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-#include "arm_compute/core/Types.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for performing a L2 normalize on a given axis given the square sum of it in this axis */
-class CLL2NormalizeLayerKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLL2NormalizeLayerKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLL2NormalizeLayerKernel(const CLL2NormalizeLayerKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLL2NormalizeLayerKernel &operator=(const CLL2NormalizeLayerKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLL2NormalizeLayerKernel(CLL2NormalizeLayerKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLL2NormalizeLayerKernel &operator=(CLL2NormalizeLayerKernel &&) = default;
-    /** Default destructor */
-    ~CLL2NormalizeLayerKernel() = default;
-
-    /** Set the input and output tensors.
-     *
-     * @param[in]  input   Source tensor. Data types supported: F16/F32. Data layouts supported: NCHW/NHWC.
-     * @param[in]  sum     Sum values tensor. Data types supported: same as @p input.
-     *                     Sum will have the same number of dimensions as input.
-     * @param[out] output  Destination tensor. Data types and data layouts supported: Same as @p input.
-     *                     Output will have the same number of dimensions as input.
-     * @param[in]  axis    Axis along which to reduce. Negative values wrap around. Maximum supported actual reduction axis : 2
-     * @param[in]  epsilon Lower bound value for the normalization.
-     */
-    void configure(const ICLTensor *input, const ICLTensor *sum, ICLTensor *output, int axis, float epsilon);
-    /** Set the input and output tensors.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Source tensor. Data types supported: F16/F32. Data layouts supported: NCHW/NHWC.
-     * @param[in]  sum             Sum values tensor. Data types supported: same as @p input.
-     *                             Sum will have the same number of dimensions as input.
-     * @param[out] output          Destination tensor. Data types and data layouts supported: Same as @p input.
-     *                             Output will have the same number of dimensions as input.
-     * @param[in]  axis            Axis along which to reduce. Negative values wrap around. Maximum supported actual reduction axis : 2
-     * @param[in]  epsilon         Lower bound value for the normalization.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, const ICLTensor *sum, ICLTensor *output, int axis, float epsilon);
-
-    /** Static function to check if given info will lead to a valid configuration of @ref CLL2NormalizeLayerKernel.
-     *
-     * @param[in] input   Source tensor info. Data types supported: F16/F32. Data layouts supported: NCHW/NHWC.
-     * @param[in] sum     Sum values tensor info. Data types supported: same as @p input.
-     *                    Sum will have the same number of dimensions as input.
-     * @param[in] output  Destination tensor info. Data types and data layouts supported: Same as @p input.
-     *                    Output will have the same number of dimensions as input.
-     * @param[in] axis    Axis along which to reduce. Negative values wrap around. Maximum supported actual reduction axis : 2
-     * @param[in] epsilon Lower bound value for the normalization.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *sum, const ITensorInfo *output, int axis, float epsilon);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;
-    const ICLTensor *_sum;
-    ICLTensor       *_output;
-    unsigned int     _actual_axis;
-    float            _epsilon;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLL2NORMALIZELAYERKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLLKTrackerKernel.h b/arm_compute/core/CL/kernels/CLLKTrackerKernel.h
deleted file mode 100644
index fdc2ef8333..0000000000
--- a/arm_compute/core/CL/kernels/CLLKTrackerKernel.h
+++ /dev/null
@@ -1,240 +0,0 @@
-/*
- * Copyright (c) 2017-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLLKTRACKERKERNEL_H
-#define ARM_COMPUTE_CLLKTRACKERKERNEL_H
-
-#include "arm_compute/core/CL/ICLArray.h"
-#include "arm_compute/core/CL/ICLKernel.h"
-#include "arm_compute/core/Types.h"
-
-#include <cstddef>
-#include <cstdint>
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Internal keypoint structure for Lucas-Kanade Optical Flow */
-struct CLLKInternalKeypoint
-{
-    float x{ 0.f };               /**< x coordinate of the keypoint */
-    float y{ 0.f };               /**< y coordinate of the keypoint */
-    float tracking_status{ 0.f }; /**< the tracking status of the keypoint */
-    float dummy{ 0.f };           /**< Dummy field, to make sure the data structure 128-bit align, so that GPU can use vload4 */
-};
-
-/** Structure for storing Spatial Gradient Matrix and the minimum eigenvalue for each keypoint */
-struct CLCoefficientTable
-{
-    float A11;     /**< iA11 * FLT_SCALE */
-    float A12;     /**< iA11 * FLT_SCALE */
-    float A22;     /**< iA11 * FLT_SCALE */
-    float min_eig; /**< Minimum eigenvalue */
-};
-
-/** Structure for storing ival, ixval and iyval for each point inside the window */
-struct CLOldValue
-{
-    int16_t ival;  /**< ival extracts from old image */
-    int16_t ixval; /**< ixval extracts from scharr Gx image */
-    int16_t iyval; /**< iyval extracts from scharr Gy image */
-    int16_t dummy; /**< Dummy field, to make sure the data structure 128-bit align, so that GPU can use vload4 */
-};
-
-/** Interface for OpenCL Array of Internal Key Points. */
-using ICLLKInternalKeypointArray = ICLArray<CLLKInternalKeypoint>;
-/** Interface for OpenCL Array of Coefficient Tables. */
-using ICLCoefficientTableArray = ICLArray<CLCoefficientTable>;
-/** Interface for OpenCL Array of Old Values. */
-using ICLOldValArray = ICLArray<CLOldValue>;
-
-/** Interface to run the initialization step of LKTracker */
-class CLLKTrackerInitKernel : public ICLKernel
-{
-public:
-    /** Initialise the kernel input and output
-     *
-     * @param[in]  old_points           Pointer to the @ref ICLKeyPointArray storing old key points
-     * @param[in]  new_points_estimates Pointer to the @ref ICLKeyPointArray storing new estimates key points
-     * @param[out] old_points_internal  Pointer to the array of internal @ref CLLKInternalKeypoint old points
-     * @param[out] new_points_internal  Pointer to the array of internal @ref CLLKInternalKeypoint new points
-     * @param[in]  use_initial_estimate The flag to indicate whether the initial estimated position should be used
-     * @param[in]  level                The pyramid level
-     * @param[in]  num_levels           The number of pyramid levels
-     * @param[in]  pyramid_scale        Scale factor used for generating the pyramid
-     */
-    void configure(const ICLKeyPointArray *old_points, const ICLKeyPointArray *new_points_estimates,
-                   ICLLKInternalKeypointArray *old_points_internal, ICLLKInternalKeypointArray *new_points_internal,
-                   bool use_initial_estimate, size_t level, size_t num_levels, float pyramid_scale);
-    /** Initialise the kernel input and output
-     *
-     * @param[in]  compile_context      The compile context to be used.
-     * @param[in]  old_points           Pointer to the @ref ICLKeyPointArray storing old key points
-     * @param[in]  new_points_estimates Pointer to the @ref ICLKeyPointArray storing new estimates key points
-     * @param[out] old_points_internal  Pointer to the array of internal @ref CLLKInternalKeypoint old points
-     * @param[out] new_points_internal  Pointer to the array of internal @ref CLLKInternalKeypoint new points
-     * @param[in]  use_initial_estimate The flag to indicate whether the initial estimated position should be used
-     * @param[in]  level                The pyramid level
-     * @param[in]  num_levels           The number of pyramid levels
-     * @param[in]  pyramid_scale        Scale factor used for generating the pyramid
-     */
-    void configure(const CLCompileContext &compile_context, const ICLKeyPointArray *old_points, const ICLKeyPointArray *new_points_estimates,
-                   ICLLKInternalKeypointArray *old_points_internal, ICLLKInternalKeypointArray *new_points_internal,
-                   bool use_initial_estimate, size_t level, size_t num_levels, float pyramid_scale);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-};
-
-/** Interface to run the finalize step of LKTracker, where it truncates the coordinates stored in new_points array */
-class CLLKTrackerFinalizeKernel : public ICLKernel
-{
-public:
-    /** Initialise the kernel input and output
-     *
-     * @param[in]  new_points_internal Pointer to the array of internal @ref CLLKInternalKeypoint new points
-     * @param[out] new_points          Pointer to the @ref ICLKeyPointArray storing new key points
-     */
-    void configure(ICLLKInternalKeypointArray *new_points_internal, ICLKeyPointArray *new_points);
-    /** Initialise the kernel input and output
-     *
-     * @param[in]  compile_context     The compile context to be used.
-     * @param[in]  new_points_internal Pointer to the array of internal @ref CLLKInternalKeypoint new points
-     * @param[out] new_points          Pointer to the @ref ICLKeyPointArray storing new key points
-     */
-    void configure(const CLCompileContext &compile_context, ICLLKInternalKeypointArray *new_points_internal, ICLKeyPointArray *new_points);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-};
-
-/** Interface to run the first stage of LKTracker, where A11, A12, A22, min_eig, ival, ixval and iyval are computed */
-class CLLKTrackerStage0Kernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLLKTrackerStage0Kernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLLKTrackerStage0Kernel(const CLLKTrackerStage0Kernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLLKTrackerStage0Kernel &operator=(const CLLKTrackerStage0Kernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLLKTrackerStage0Kernel(CLLKTrackerStage0Kernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLLKTrackerStage0Kernel &operator=(CLLKTrackerStage0Kernel &&) = default;
-    /** Initialise the kernel input and output
-     *
-     * @param[in]      old_input           Pointer to the input old tensor. Data types supported: U8
-     * @param[in]      old_scharr_gx       Pointer to the input scharr X tensor. Data types supported: S16
-     * @param[in]      old_scharr_gy       Pointer to the input scharr Y tensor. Data types supported: S16
-     * @param[in]      old_points_internal Pointer to the array of CLLKInternalKeypoint old points
-     * @param[in, out] new_points_internal Pointer to the array of CLLKInternalKeypoint new points
-     * @param[out]     coeff_table         Pointer to the array holding the Spatial Gradient coefficients
-     * @param[out]     old_ival            Pointer to the array holding internal values
-     * @param[in]      window_dimension    The size of the window on which to perform the algorithm
-     * @param[in]      level               The pyramid level
-     */
-    void configure(const ICLTensor *old_input, const ICLTensor *old_scharr_gx, const ICLTensor *old_scharr_gy,
-                   ICLLKInternalKeypointArray *old_points_internal, ICLLKInternalKeypointArray *new_points_internal,
-                   ICLCoefficientTableArray *coeff_table, ICLOldValArray *old_ival,
-                   size_t window_dimension, size_t level);
-    /** Initialise the kernel input and output
-     *
-     * @param[in]      compile_context     The compile context to be used.
-     * @param[in]      old_input           Pointer to the input old tensor. Data types supported: U8
-     * @param[in]      old_scharr_gx       Pointer to the input scharr X tensor. Data types supported: S16
-     * @param[in]      old_scharr_gy       Pointer to the input scharr Y tensor. Data types supported: S16
-     * @param[in]      old_points_internal Pointer to the array of CLLKInternalKeypoint old points
-     * @param[in, out] new_points_internal Pointer to the array of CLLKInternalKeypoint new points
-     * @param[out]     coeff_table         Pointer to the array holding the Spatial Gradient coefficients
-     * @param[out]     old_ival            Pointer to the array holding internal values
-     * @param[in]      window_dimension    The size of the window on which to perform the algorithm
-     * @param[in]      level               The pyramid level
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *old_input, const ICLTensor *old_scharr_gx, const ICLTensor *old_scharr_gy,
-                   ICLLKInternalKeypointArray *old_points_internal, ICLLKInternalKeypointArray *new_points_internal,
-                   ICLCoefficientTableArray *coeff_table, ICLOldValArray *old_ival,
-                   size_t window_dimension, size_t level);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_old_input;
-    const ICLTensor *_old_scharr_gx;
-    const ICLTensor *_old_scharr_gy;
-};
-
-/** Interface to run the second stage of LKTracker, where the motion vectors of the given points are computed */
-class CLLKTrackerStage1Kernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLLKTrackerStage1Kernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLLKTrackerStage1Kernel(const CLLKTrackerStage1Kernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLLKTrackerStage1Kernel &operator=(const CLLKTrackerStage1Kernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLLKTrackerStage1Kernel(CLLKTrackerStage1Kernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLLKTrackerStage1Kernel &operator=(CLLKTrackerStage1Kernel &&) = default;
-    /** Initialise the kernel input and output
-     *
-     * @param[in]      new_input           Pointer to the input new tensor. Data types supported: U8
-     * @param[in, out] new_points_internal Pointer to the array of CLLKInternalKeypoint for new points
-     * @param[in]      coeff_table         Pointer to the array holding the Spatial Gradient coefficients
-     * @param[in]      old_ival            Pointer to the array holding internal values
-     * @param[in]      termination         The criteria to terminate the search of each keypoint.
-     * @param[in]      epsilon             The error for terminating the algorithm
-     * @param[in]      num_iterations      The maximum number of iterations before terminating the algorithm
-     * @param[in]      window_dimension    The size of the window on which to perform the algorithm
-     * @param[in]      level               The pyramid level
-     */
-    void configure(const ICLTensor *new_input, ICLLKInternalKeypointArray *new_points_internal, ICLCoefficientTableArray *coeff_table, ICLOldValArray *old_ival,
-                   Termination termination, float epsilon, size_t num_iterations, size_t window_dimension, size_t level);
-    /** Initialise the kernel input and output
-     *
-     * @param[in]      compile_context     The compile context to be used.
-     * @param[in]      new_input           Pointer to the input new tensor. Data types supported: U8
-     * @param[in, out] new_points_internal Pointer to the array of CLLKInternalKeypoint for new points
-     * @param[in]      coeff_table         Pointer to the array holding the Spatial Gradient coefficients
-     * @param[in]      old_ival            Pointer to the array holding internal values
-     * @param[in]      termination         The criteria to terminate the search of each keypoint.
-     * @param[in]      epsilon             The error for terminating the algorithm
-     * @param[in]      num_iterations      The maximum number of iterations before terminating the algorithm
-     * @param[in]      window_dimension    The size of the window on which to perform the algorithm
-     * @param[in]      level               The pyramid level
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *new_input, ICLLKInternalKeypointArray *new_points_internal, ICLCoefficientTableArray *coeff_table, ICLOldValArray *old_ival,
-                   Termination termination, float epsilon, size_t num_iterations, size_t window_dimension, size_t level);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_new_input;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLLKTRACKERKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLLocallyConnectedMatrixMultiplyKernel.h b/arm_compute/core/CL/kernels/CLLocallyConnectedMatrixMultiplyKernel.h
deleted file mode 100644
index d5653f83ea..0000000000
--- a/arm_compute/core/CL/kernels/CLLocallyConnectedMatrixMultiplyKernel.h
+++ /dev/null
@@ -1,85 +0,0 @@
-/*
- * Copyright (c) 2017-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLLOCALLYCONNECTEDMATRIXMULTIPLYKERNEL_H
-#define ARM_COMPUTE_CLLOCALLYCONNECTEDMATRIXMULTIPLYKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** OpenCL kernel to multiply each row of first tensor with low 2 dimensions of second tensor.
- *
- * @attention The second input tensor must have at least 2 dimensions (matrix)
- *
- */
-class CLLocallyConnectedMatrixMultiplyKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLLocallyConnectedMatrixMultiplyKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLLocallyConnectedMatrixMultiplyKernel(const CLLocallyConnectedMatrixMultiplyKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLLocallyConnectedMatrixMultiplyKernel &operator=(const CLLocallyConnectedMatrixMultiplyKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLLocallyConnectedMatrixMultiplyKernel(CLLocallyConnectedMatrixMultiplyKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLLocallyConnectedMatrixMultiplyKernel &operator=(CLLocallyConnectedMatrixMultiplyKernel &&) = default;
-    /** Initialise the kernel's input, output and alpha
-     *
-     * @param[in]  input0 First input tensor. Data types supported: F32
-     * @param[in]  input1 Second input tensor. Data type supported: same as @p input0
-     * @param[out] output Output tensor to store the result. Data type supported: same as @p input0
-     */
-    void configure(const ICLTensor *input0, const ICLTensor *input1, ICLTensor *output);
-    /** Initialise the kernel's input, output and alpha
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input0          First input tensor. Data types supported: F32
-     * @param[in]  input1          Second input tensor. Data type supported: same as @p input0
-     * @param[out] output          Output tensor to store the result. Data type supported: same as @p input0
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input0, const ICLTensor *input1, ICLTensor *output);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLLocallyConnectedMatrixMultiplyKernel
-     *
-     * @param[in] input0 First input tensor info. Data types supported: F32
-     * @param[in] input1 Second input tensor info. Data type supported: same as @p input0
-     * @param[in] output Output tensor info. Data type supported: same as @p input0
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input0, const ITensorInfo *input1, const ITensorInfo *output);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input0;
-    const ICLTensor *_input1;
-    ICLTensor       *_output;
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLLOCALLYCONNECTEDMATRIXMULTIPLYKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLMagnitudePhaseKernel.h b/arm_compute/core/CL/kernels/CLMagnitudePhaseKernel.h
deleted file mode 100644
index a741b1745a..0000000000
--- a/arm_compute/core/CL/kernels/CLMagnitudePhaseKernel.h
+++ /dev/null
@@ -1,90 +0,0 @@
-/*
- * Copyright (c) 2016-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLMAGNITUDEPHASEKERNEL_H
-#define ARM_COMPUTE_CLMAGNITUDEPHASEKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-#include "arm_compute/core/Types.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Template interface for the kernel to compute magnitude and phase.
- *
- */
-class CLMagnitudePhaseKernel : public ICLKernel
-{
-public:
-    /** Default constructor. */
-    CLMagnitudePhaseKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLMagnitudePhaseKernel(const CLMagnitudePhaseKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLMagnitudePhaseKernel &operator=(const CLMagnitudePhaseKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLMagnitudePhaseKernel(CLMagnitudePhaseKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLMagnitudePhaseKernel &operator=(CLMagnitudePhaseKernel &&) = default;
-    /** Initialise the kernel's input, output.
-     *
-     * @note At least one of output1 or output2 must be set.
-     *
-     * @param[in]  gx         The input gradient X tensor. Data types supported: S16/S32.
-     * @param[in]  gy         The input gradient Y tensor. Data types supported: S16/S32.
-     * @param[out] magnitude  (Optional) The output tensor - Magnitude. Data types supported: S16/S32.
-     * @param[out] phase      (Optional) The output tensor - Phase. Data types supported: U8.
-     * @param[in]  mag_type   (Optional) Magnitude calculation type. Default: L2NORM.
-     * @param[in]  phase_type (Optional) Phase calculation type. Default: SIGNED.
-     */
-    void configure(const ICLTensor *gx, const ICLTensor *gy, ICLTensor *magnitude, ICLTensor *phase,
-                   MagnitudeType mag_type = MagnitudeType::L2NORM, PhaseType phase_type = PhaseType::SIGNED);
-    /** Initialise the kernel's input, output.
-     *
-     * @note At least one of output1 or output2 must be set.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  gx              The input gradient X tensor. Data types supported: S16/S32.
-     * @param[in]  gy              The input gradient Y tensor. Data types supported: S16/S32.
-     * @param[out] magnitude       (Optional) The output tensor - Magnitude. Data types supported: S16/S32.
-     * @param[out] phase           (Optional) The output tensor - Phase. Data types supported: U8.
-     * @param[in]  mag_type        (Optional) Magnitude calculation type. Default: L2NORM.
-     * @param[in]  phase_type      (Optional) Phase calculation type. Default: SIGNED.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *gx, const ICLTensor *gy, ICLTensor *magnitude, ICLTensor *phase,
-                   MagnitudeType mag_type = MagnitudeType::L2NORM, PhaseType phase_type = PhaseType::SIGNED);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_gx;        /**< Input gradient X. */
-    const ICLTensor *_gy;        /**< Input gradient Y. */
-    ICLTensor       *_magnitude; /**< Output - Magnitude. */
-    ICLTensor       *_phase;     /**< Output - Phase. */
-    bool             _run_mag;   /**< Calculate magnitude ? */
-    bool             _run_phase; /**< Calculate phase ? */
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLMAGNITUDEPHASEKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLMaxUnpoolingLayerKernel.h b/arm_compute/core/CL/kernels/CLMaxUnpoolingLayerKernel.h
deleted file mode 100644
index 9d51f6b59c..0000000000
--- a/arm_compute/core/CL/kernels/CLMaxUnpoolingLayerKernel.h
+++ /dev/null
@@ -1,86 +0,0 @@
-/*
- * Copyright (c) 2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLMAXUNPOOLINGLAYERKERNEL_H
-#define ARM_COMPUTE_CLMAXUNPOOLINGLAYERKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the pooling layer kernel */
-class CLMaxUnpoolingLayerKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLMaxUnpoolingLayerKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLMaxUnpoolingLayerKernel(const CLMaxUnpoolingLayerKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLMaxUnpoolingLayerKernel &operator=(const CLMaxUnpoolingLayerKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLMaxUnpoolingLayerKernel(CLMaxUnpoolingLayerKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLMaxUnpoolingLayerKernel &operator=(CLMaxUnpoolingLayerKernel &&) = default;
-    /** Default destructor */
-    ~CLMaxUnpoolingLayerKernel() = default;
-    /** Set the input and output tensors.
-     *
-     * @note Output shape must be equal to the shape of the original input to pool.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Source tensor. Data types supported: QASYMM8/QASYMM8_SIGNED/F16/F32.
-     * @param[in]  indices         Tensor containing the offset to store the input elements in the output tensor.
-     *                             @ref CLPoolingLayerKernel with indices should precede this function in order to
-     *                             properly reconstruct the output tensor.
-     *                             The tensor shape of this tensor has to be equal to the input tensor shape. Data type supported: U32.
-     * @param[out] output          Destination tensor. Data types supported: Same as @p input.
-     * @param[in]  pool_info       Contains pooling operation information described in @ref PoolingLayerInfo.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, const ICLTensor *indices, ICLTensor *output, const PoolingLayerInfo &pool_info);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLMaxUnpoolingLayerKernel
-     *
-     * @param[in] input     Source tensor info. Data types supported: QASYMM8/QASYMM8_SIGNED/F16/F32.
-     * @param[in] output    Destination tensor info. Data types supported: Same as @p input.
-     * @param[in] indices   TensorInfo associated to the tensor containing the offset to store the input elements in the output tensor.
-     *                      @ref CLPoolingLayerKernel with indices should precede this function in order to
-     *                      properly reconstruct the output tensor.
-     *                      The tensor shape of this tensor has to be equal to the input tensor shape. Data type supported: U32.
-     * @param[in] pool_info Contains pooling operation information described in @ref PoolingLayerInfo.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *indices, const ITensorInfo *output, const PoolingLayerInfo &pool_info);
-
-    // Inherited methods overridden
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;
-    ICLTensor       *_output;
-    const ICLTensor *_indices;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLMAXUNPOOLINGLAYERKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLMeanStdDevKernel.h b/arm_compute/core/CL/kernels/CLMeanStdDevKernel.h
deleted file mode 100644
index 2a5a5f2e33..0000000000
--- a/arm_compute/core/CL/kernels/CLMeanStdDevKernel.h
+++ /dev/null
@@ -1,98 +0,0 @@
-/*
- * Copyright (c) 2016-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLMEANSTDDEVKERNEL_H
-#define ARM_COMPUTE_CLMEANSTDDEVKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace cl
-{
-class Buffer;
-}
-
-namespace arm_compute
-{
-class ICLTensor;
-using ICLImage = ICLTensor;
-
-/** Interface for the kernel to calculate mean and standard deviation of input image pixels. */
-class CLMeanStdDevKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLMeanStdDevKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLMeanStdDevKernel(const CLMeanStdDevKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLMeanStdDevKernel &operator=(const CLMeanStdDevKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLMeanStdDevKernel(CLMeanStdDevKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLMeanStdDevKernel &operator=(CLMeanStdDevKernel &&) = default;
-    /** Initialise the kernel's input and outputs.
-     *
-     * @param[in]  input              Input image. Data types supported: U8.
-     * @param[out] mean               Input average pixel value.
-     * @param[out] global_sum         Keeps global sum of pixel values (Buffer size: 1 cl_ulong).
-     * @param[out] stddev             (Optional) Output standard deviation of pixel values.
-     * @param[out] global_sum_squared (Optional if stddev is not set, required if stddev is set) Keeps global sum of squared pixel values (Buffer size: 1 cl_ulong).
-     */
-    void configure(const ICLImage *input, float *mean, cl::Buffer *global_sum, float *stddev = nullptr, cl::Buffer *global_sum_squared = nullptr);
-    /** Initialise the kernel's input and outputs.
-     *
-     * @param[in]  compile_context    The compile context to be used.
-     * @param[in]  input              Input image. Data types supported: U8.
-     * @param[out] mean               Input average pixel value.
-     * @param[out] global_sum         Keeps global sum of pixel values (Buffer size: 1 cl_ulong).
-     * @param[out] stddev             (Optional) Output standard deviation of pixel values.
-     * @param[out] global_sum_squared (Optional if stddev is not set, required if stddev is set) Keeps global sum of squared pixel values (Buffer size: 1 cl_ulong).
-     */
-    void configure(const CLCompileContext &compile_context, const ICLImage *input, float *mean, cl::Buffer *global_sum, float *stddev = nullptr, cl::Buffer *global_sum_squared = nullptr);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLMeanStdDevKernel.
-     *
-     * @param[in] input              Input image info. Data types supported: U8.
-     * @param[in] mean               Input average pixel value.
-     * @param[in] global_sum         Keeps global sum of pixel values.
-     * @param[in] stddev             (Optional) Output standard deviation of pixel values.
-     * @param[in] global_sum_squared (Optional if stddev is not set, required if stddev is set) Keeps global sum of squared pixel values.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, float *mean, cl::Buffer *global_sum, float *stddev = nullptr, cl::Buffer *global_sum_squared = nullptr);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-    BorderSize border_size() const override;
-
-private:
-    const ICLImage *_input;
-    float          *_mean;
-    float          *_stddev;
-    cl::Buffer     *_global_sum;
-    cl::Buffer     *_global_sum_squared;
-    BorderSize      _border_size;
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLMEANSTDDEVKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLMeanStdDevNormalizationKernel.h b/arm_compute/core/CL/kernels/CLMeanStdDevNormalizationKernel.h
deleted file mode 100644
index ff5e9ab0f7..0000000000
--- a/arm_compute/core/CL/kernels/CLMeanStdDevNormalizationKernel.h
+++ /dev/null
@@ -1,90 +0,0 @@
-/*
- * Copyright (c) 2019-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLMEANSTDDEVNORMALIZATIONKERNEL_H
-#define ARM_COMPUTE_CLMEANSTDDEVNORMALIZATIONKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the kernel to normalize the input 2D tensor across the first dimension with respect to mean and standard deviation of the same dimension. */
-class CLMeanStdDevNormalizationKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLMeanStdDevNormalizationKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLMeanStdDevNormalizationKernel(const CLMeanStdDevNormalizationKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLMeanStdDevNormalizationKernel &operator=(const CLMeanStdDevNormalizationKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLMeanStdDevNormalizationKernel(CLMeanStdDevNormalizationKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLMeanStdDevNormalizationKernel &operator=(CLMeanStdDevNormalizationKernel &&) = default;
-    /** Default destructor */
-    ~CLMeanStdDevNormalizationKernel() = default;
-    /** Initialise the kernel's input and outputs.
-     *
-     * @note If the output tensor is a nullptr, the normalization will be performed in-place.
-     *
-     * @param[in, out] input   Source tensor with 2 dimensions. In case of @p output tensor = nullptr,
-     *                         this tensor will store the result of the normalization. Data types supported: F16/F32.
-     * @param[out]     output  (Optional) Destination tensor. It can be nullptr in case of in-place computation. Data type supported: same as @p input
-     * @param[in]      epsilon (Optional) Small float to avoid division by zero in case of zero standard deviation. Defaults to 1e-8.
-     */
-    void configure(ICLTensor *input, ICLTensor *output = nullptr, float epsilon = 1e-8f);
-    /** Initialise the kernel's input and outputs.
-     *
-     * @note If the output tensor is a nullptr, the normalization will be performed in-place.
-     *
-     * @param[in]      compile_context The compile context to be used.
-     * @param[in, out] input           Source tensor with 2 dimensions. In case of @p output tensor = nullptr,
-     *                                 this tensor will store the result of the normalization. Data types supported: F16/F32.
-     * @param[out]     output          (Optional) Destination tensor. It can be nullptr in case of in-place computation. Data type supported: same as @p input
-     * @param[in]      epsilon         (Optional) Small float to avoid division by zero in case of zero standard deviation. Defaults to 1e-8.
-     */
-    void configure(const CLCompileContext &compile_context, ICLTensor *input, ICLTensor *output = nullptr, float epsilon = 1e-8f);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLMeanStdDevNormalizationKernel
-     *
-     * @param[in] input   Source tensor info with 2 dimensions. In case of @p output tensor info = nullptr,
-     *                    this tensor will store the result of the normalization. Data types supported: F16/F32.
-     * @param[in] output  (Optional) Destination tensor info. It can be nullptr in case of in-place computation. Data type supported: same as @p input
-     * @param[in] epsilon (Optional) Small float to avoid division by zero in case of zero standard deviation. Defaults to 1e-8.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output = nullptr, float epsilon = 1e-8f);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    ICLTensor *_input;
-    ICLTensor *_output;
-    bool       _run_in_place;
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLMEANSTDDEVNORMALIZATIONKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLMedian3x3Kernel.h b/arm_compute/core/CL/kernels/CLMedian3x3Kernel.h
deleted file mode 100644
index ccb475360f..0000000000
--- a/arm_compute/core/CL/kernels/CLMedian3x3Kernel.h
+++ /dev/null
@@ -1,59 +0,0 @@
-/*
- * Copyright (c) 2016-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLMEDIAN3X3KERNEL_H
-#define ARM_COMPUTE_CLMEDIAN3X3KERNEL_H
-
-#include "arm_compute/core/CL/ICLSimple2DKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the median 3x3 filter kernel.
- *
- */
-class CLMedian3x3Kernel : public ICLSimple2DKernel
-{
-public:
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in]  input            An input tensor. Data types supported: U8
-     * @param[out] output           The output tensor. Data types supported: U8.
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, bool border_undefined);
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in]  compile_context  The compile context to be used.
-     * @param[in]  input            An input tensor. Data types supported: U8
-     * @param[out] output           The output tensor. Data types supported: U8.
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, bool border_undefined);
-
-    // Inherited methods overridden:
-    BorderSize border_size() const override;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLMEDIAN3X3KERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLMemsetKernel.h b/arm_compute/core/CL/kernels/CLMemsetKernel.h
deleted file mode 100644
index 5bda480306..0000000000
--- a/arm_compute/core/CL/kernels/CLMemsetKernel.h
+++ /dev/null
@@ -1,85 +0,0 @@
-/*
- * Copyright (c) 2018-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLMEMSETKERNEL_H
-#define ARM_COMPUTE_CLMEMSETKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-#include "arm_compute/core/PixelValue.h"
-#include "arm_compute/core/Types.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for filling the planes of a tensor */
-class CLMemsetKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLMemsetKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLMemsetKernel(const CLMemsetKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLMemsetKernel &operator=(const CLMemsetKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLMemsetKernel(CLMemsetKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLMemsetKernel &operator=(CLMemsetKernel &&) = default;
-    /** Default destructor */
-    ~CLMemsetKernel() = default;
-
-    /** Initialise the kernel's tensor and filling value
-     *
-     * @param[in,out] tensor         Input tensor to fill. Supported data types: All.
-     * @param[in]     constant_value The value used to fill the planes of the tensor
-     * @param[in]     window         Window to be used in case setting only part of a tensor. Default is nullptr.
-     */
-    void configure(ICLTensor *tensor, const PixelValue &constant_value, Window *window = nullptr);
-    /** Initialise the kernel's tensor and filling value
-     *
-     * @param[in]     compile_context The compile context to be used.
-     * @param[in,out] tensor          Input tensor to fill. Supported data types: All.
-     * @param[in]     constant_value  The value used to fill the planes of the tensor
-     * @param[in]     window          Window to be used in case setting only part of a tensor. Default is nullptr.
-     */
-    void configure(const CLCompileContext &compile_context, ICLTensor *tensor, const PixelValue &constant_value, Window *window = nullptr);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLMemsetKernel
-     *
-     * @param[in] tensor         Source tensor info. Data types supported: All.
-     * @param[in] constant_value The value used to fill the planes of the tensor
-     * @param[in] window         Window to be used in case setting only part of a tensor. Default is nullptr.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *tensor, const PixelValue &constant_value, Window *window = nullptr);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    ICLTensor *_tensor;
-    Window     _full_window;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLMEMSETRKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLMinMaxLayerKernel.h b/arm_compute/core/CL/kernels/CLMinMaxLayerKernel.h
deleted file mode 100644
index a693cfdb27..0000000000
--- a/arm_compute/core/CL/kernels/CLMinMaxLayerKernel.h
+++ /dev/null
@@ -1,87 +0,0 @@
-/*
- * Copyright (c) 2017-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLMINMAXLAYERKERNEL_H
-#define ARM_COMPUTE_CLMINMAXLAYERKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the kernel to perform min max search on a 3D tensor.
- */
-class CLMinMaxLayerKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLMinMaxLayerKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLMinMaxLayerKernel(const CLMinMaxLayerKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLMinMaxLayerKernel &operator=(const CLMinMaxLayerKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLMinMaxLayerKernel(CLMinMaxLayerKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLMinMaxLayerKernel &operator=(CLMinMaxLayerKernel &&) = default;
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in]  input  Input tensor with at least 3 dimensions. The dimensions over the third will be interpreted as batches.Data types supported: F32.
-     * @param[out] output Output tensor with shape [2, batches, ...] which stores the minimum and maximum values for each 3D input tensor.
-     *                    The dimensions over the second must match the batched dimensions of the input tensor. Data types supported: F32.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output);
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Input tensor with at least 3 dimensions. The dimensions over the third will be interpreted as batches.Data types supported: F32.
-     * @param[out] output          Output tensor with shape [2, batches, ...] which stores the minimum and maximum values for each 3D input tensor.
-     *                    The dimensions over the second must match the batched dimensions of the input tensor. Data types supported: F32.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLMinMaxLayerKernel
-     *
-     * @param[in] input  Input tensor info.  Data types supported: F32.
-     * @param[in] output Output tensor info with shape [2, batches, ...] which stores the minimum and maximum values for each 3D input tensor.
-     *                   The dimensions over the second must match the batched dimensions of the input tensor. Data types supported: F32.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output);
-
-    /** Resets global minimum and maximum
-     *
-     * @param[in,out] queue Command queue on which to map and unmap the min_max tensor
-     */
-    void reset(cl::CommandQueue &queue);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;
-    ICLTensor       *_output;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLMINMAXLAYERKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLMinMaxLocationKernel.h b/arm_compute/core/CL/kernels/CLMinMaxLocationKernel.h
deleted file mode 100644
index fbcf69752c..0000000000
--- a/arm_compute/core/CL/kernels/CLMinMaxLocationKernel.h
+++ /dev/null
@@ -1,124 +0,0 @@
-/*
- * Copyright (c) 2017-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLMINMAXLOCATIONKERNEL_H
-#define ARM_COMPUTE_CLMINMAXLOCATIONKERNEL_H
-
-#include "arm_compute/core/CL/ICLArray.h"
-#include "arm_compute/core/CL/ICLKernel.h"
-
-#include <array>
-
-namespace arm_compute
-{
-class ICLTensor;
-using ICLImage = ICLTensor;
-
-/** Interface for the kernel to perform min max search on an image.
- */
-class CLMinMaxKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLMinMaxKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLMinMaxKernel(const CLMinMaxKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLMinMaxKernel &operator=(const CLMinMaxKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLMinMaxKernel(CLMinMaxKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLMinMaxKernel &operator=(CLMinMaxKernel &&) = default;
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in]  input   Input Image. Data types supported: U8/S16/F32.
-     * @param[out] min_max Buffer of 2 elements to store the min value at position 0 and the max value at position 1. Data type supported: S32 if input type is U8/S16, F32 if input type is F32.
-     */
-    void configure(const ICLImage *input, cl::Buffer *min_max);
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Input Image. Data types supported: U8/S16/F32.
-     * @param[out] min_max         Buffer of 2 elements to store the min value at position 0 and the max value at position 1. Data type supported: S32 if input type is U8/S16, F32 if input type is F32.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLImage *input, cl::Buffer *min_max);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;               /**< Input image. */
-    cl::Buffer      *_min_max;             /**< Minimum/maximum value. */
-    std::array<int, 2> _data_type_max_min; /**< Maximum and minimum data type value respectively. */
-};
-
-/** Interface for the kernel to find min max locations of an image.
- */
-class CLMinMaxLocationKernel : public ICLKernel
-{
-public:
-    /** Constructor */
-    CLMinMaxLocationKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLMinMaxLocationKernel(const CLMinMaxLocationKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLMinMaxLocationKernel &operator=(const CLMinMaxLocationKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLMinMaxLocationKernel(CLMinMaxLocationKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLMinMaxLocationKernel &operator=(CLMinMaxLocationKernel &&) = default;
-    /** Initialise the kernel's input and outputs.
-     *
-     * @note When locations of min and max occurrences are requested, the reported number of locations is limited to the given array size.
-     *
-     * @param[in]  input         Input image. Data types supported: U8/S16/F32.
-     * @param[out] min_max       Buffer of 2 elements to store the min value at position 0 and the max value at position 1. Data type supported: S32 if input type is U8/S16, F32 if input type is F32.
-     * @param[out] min_max_count Buffer of 2 elements to store the min value occurrences at position 0 and the max value occurrences at position 1. Data type supported: S32
-     * @param[out] min_loc       (Optional) Array of Coordinates2D used to store minimum value locations.
-     * @param[out] max_loc       (Optional) Array of Coordinates2D used to store maximum value locations.
-     */
-    void configure(const ICLImage *input, cl::Buffer *min_max, cl::Buffer *min_max_count,
-                   ICLCoordinates2DArray *min_loc = nullptr, ICLCoordinates2DArray *max_loc = nullptr);
-    /** Initialise the kernel's input and outputs.
-     *
-     * @note When locations of min and max occurrences are requested, the reported number of locations is limited to the given array size.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Input image. Data types supported: U8/S16/F32.
-     * @param[out] min_max         Buffer of 2 elements to store the min value at position 0 and the max value at position 1. Data type supported: S32 if input type is U8/S16, F32 if input type is F32.
-     * @param[out] min_max_count   Buffer of 2 elements to store the min value occurrences at position 0 and the max value occurrences at position 1. Data type supported: S32
-     * @param[out] min_loc         (Optional) Array of Coordinates2D used to store minimum value locations.
-     * @param[out] max_loc         (Optional) Array of Coordinates2D used to store maximum value locations.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLImage *input, cl::Buffer *min_max, cl::Buffer *min_max_count,
-                   ICLCoordinates2DArray *min_loc = nullptr, ICLCoordinates2DArray *max_loc = nullptr);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLImage *_input;         /**< Input image. */
-    cl::Buffer     *_min_max_count; /**< Minimum/maximum value occurrences. */
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLMINMAXLOCATIONKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLNonLinearFilterKernel.h b/arm_compute/core/CL/kernels/CLNonLinearFilterKernel.h
deleted file mode 100644
index cee64480b6..0000000000
--- a/arm_compute/core/CL/kernels/CLNonLinearFilterKernel.h
+++ /dev/null
@@ -1,77 +0,0 @@
-/*
- * Copyright (c) 2017-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLNONLINEARFILTERKERNEL_H
-#define ARM_COMPUTE_CLNONLINEARFILTERKERNEL_H
-
-#include "arm_compute/core/CL/ICLSimple2DKernel.h"
-#include "arm_compute/core/Types.h"
-
-#include <cstdint>
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the kernel to apply a non-linear filter */
-class CLNonLinearFilterKernel : public ICLSimple2DKernel
-{
-public:
-    /** Default constructor */
-    CLNonLinearFilterKernel();
-    /** Set the source, destination and border mode of the kernel
-     *
-     * @param[in]  input            Source tensor. Data types supported: U8
-     * @param[out] output           Destination tensor. Data types supported: U8
-     * @param[in]  function         Non linear function to perform
-     * @param[in]  mask_size        Mask size. Supported sizes: 3, 5
-     * @param[in]  pattern          Mask pattern
-     * @param[in]  mask             The given mask. Will be used only if pattern is specified to PATTERN_OTHER
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, NonLinearFilterFunction function,
-                   unsigned int mask_size, MatrixPattern pattern, const uint8_t *mask,
-                   bool border_undefined);
-    /** Set the source, destination and border mode of the kernel
-     *
-     * @param[in]  compile_context  The compile context to be used.
-     * @param[in]  input            Source tensor. Data types supported: U8
-     * @param[out] output           Destination tensor. Data types supported: U8
-     * @param[in]  function         Non linear function to perform
-     * @param[in]  mask_size        Mask size. Supported sizes: 3, 5
-     * @param[in]  pattern          Mask pattern
-     * @param[in]  mask             The given mask. Will be used only if pattern is specified to PATTERN_OTHER
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, NonLinearFilterFunction function,
-                   unsigned int mask_size, MatrixPattern pattern, const uint8_t *mask,
-                   bool border_undefined);
-
-    // Inherited methods overridden:
-    BorderSize border_size() const override;
-
-private:
-    BorderSize _border_size; /**< Border size */
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLNONLINEARFILTERKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLNonMaximaSuppression3x3Kernel.h b/arm_compute/core/CL/kernels/CLNonMaximaSuppression3x3Kernel.h
deleted file mode 100644
index d1bba4f480..0000000000
--- a/arm_compute/core/CL/kernels/CLNonMaximaSuppression3x3Kernel.h
+++ /dev/null
@@ -1,60 +0,0 @@
-/*
- * Copyright (c) 2016-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLNONMAXIMASUPPRESSION3x3KERNEL_H
-#define ARM_COMPUTE_CLNONMAXIMASUPPRESSION3x3KERNEL_H
-
-#include "arm_compute/core/CL/ICLSimple2DKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface to perform Non-Maxima suppression over a 3x3 window using OpenCL
- *
- * @note Used by @ref CLFastCorners and @ref CLHarrisCorners
- */
-class CLNonMaximaSuppression3x3Kernel : public ICLSimple2DKernel
-{
-public:
-    /** Initialise the kernel's sources, destinations and border mode.
-     *
-     * @param[in]  input            Source tensor. Data types supported: U8, F32. (Must be the same as the output tensor)
-     * @param[out] output           Destination tensor. Data types supported: U8, F32. (Must be the same as the input tensor)
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, bool border_undefined);
-    /** Initialise the kernel's sources, destinations and border mode.
-     *
-     * @param[in]  compile_context  The compile context to be used.
-     * @param[in]  input            Source tensor. Data types supported: U8, F32. (Must be the same as the output tensor)
-     * @param[out] output           Destination tensor. Data types supported: U8, F32. (Must be the same as the input tensor)
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, bool border_undefined);
-
-    // Inherited methods overridden:
-    BorderSize border_size() const override;
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLNONMAXIMASUPPRESSION3x3KERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLNormalizationLayerKernel.h b/arm_compute/core/CL/kernels/CLNormalizationLayerKernel.h
deleted file mode 100644
index 6233d28b0a..0000000000
--- a/arm_compute/core/CL/kernels/CLNormalizationLayerKernel.h
+++ /dev/null
@@ -1,90 +0,0 @@
-/*
- * Copyright (c) 2017-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLNORMALIZATIONLAYERKERNEL_H
-#define ARM_COMPUTE_CLNORMALIZATIONLAYERKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the normalization layer kernel.
- */
-class CLNormalizationLayerKernel : public ICLKernel
-{
-public:
-    /** Constructor */
-    CLNormalizationLayerKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLNormalizationLayerKernel(const CLNormalizationLayerKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLNormalizationLayerKernel &operator=(const CLNormalizationLayerKernel &) = delete;
-    /** Default Move Constructor. */
-    CLNormalizationLayerKernel(CLNormalizationLayerKernel &&) = default;
-    /** Default move assignment operator */
-    CLNormalizationLayerKernel &operator=(CLNormalizationLayerKernel &&) = default;
-    /** Set the input and output tensors.
-     *
-     * @param[in]  input     Source tensor. 3 lower dims represent a single input with dimensions [width, height, IFM],
-     *                       and an optional 4th dimension for batch of inputs. Data types supported: F16/F32. Data layouts supported: NCHW/NHWC.
-     * @param[out] output    Destination tensor. Output will have the same number of dimensions as input. Data types supported: same as @p input.
-     *                       Data layouts supported: same as @p input.
-     * @param[in]  norm_info Normalization layer information like the normalization type, normalization size and other parameters.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, NormalizationLayerInfo norm_info);
-    /** Set the input and output tensors.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Source tensor. 3 lower dims represent a single input with dimensions [width, height, IFM],
-     *                             and an optional 4th dimension for batch of inputs. Data types supported: F16/F32. Data layouts supported: NCHW/NHWC.
-     * @param[out] output          Destination tensor. Output will have the same number of dimensions as input. Data types supported: same as @p input.
-     *                             Data layouts supported: same as @p input.
-     * @param[in]  norm_info       Normalization layer information like the normalization type, normalization size and other parameters.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, NormalizationLayerInfo norm_info);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLNormalizationLayerKernel
-     *
-     * @param[in] input     Source tensor. 3 lower dims represent a single input with dimensions [width, height, IFM],
-     *                      and an optional 4th dimension for batch of inputs. Data types supported: F16/F32. Data layouts supported: NCHW/NHWC.
-     * @param[in] output    Destination tensor. Output will have the same number of dimensions as input. Data types supported: same as @p input.
-     *                      Data layouts supported: same as @p input.
-     * @param[in] norm_info Normalization layer information like the normalization type, normalization size and other parameters.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output, NormalizationLayerInfo norm_info);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-    BorderSize border_size() const override;
-
-private:
-    const ICLTensor *_input;
-    ICLTensor       *_output;
-    BorderSize       _border_size;
-    bool             _is_norm_across_width;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLNORMALIZATIONLAYERKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLNormalizePlanarYUVLayerKernel.h b/arm_compute/core/CL/kernels/CLNormalizePlanarYUVLayerKernel.h
deleted file mode 100644
index 2e2e60df0b..0000000000
--- a/arm_compute/core/CL/kernels/CLNormalizePlanarYUVLayerKernel.h
+++ /dev/null
@@ -1,94 +0,0 @@
-/*
- * Copyright (c) 2018-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLNORMALIZEPLANARYUVLAYERKERNEL_H
-#define ARM_COMPUTE_CLNORMALIZEPLANARYUVLAYERKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the NormalizePlanarYUV layer kernel. */
-class CLNormalizePlanarYUVLayerKernel : public ICLKernel
-{
-public:
-    /** Constructor */
-    CLNormalizePlanarYUVLayerKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLNormalizePlanarYUVLayerKernel(const CLNormalizePlanarYUVLayerKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLNormalizePlanarYUVLayerKernel &operator=(const CLNormalizePlanarYUVLayerKernel &) = delete;
-    /** Default Move Constructor. */
-    CLNormalizePlanarYUVLayerKernel(CLNormalizePlanarYUVLayerKernel &&) = default;
-    /** Default move assignment operator */
-    CLNormalizePlanarYUVLayerKernel &operator=(CLNormalizePlanarYUVLayerKernel &&) = default;
-    /** Default destructor */
-    ~CLNormalizePlanarYUVLayerKernel() = default;
-
-    /** Set the input and output tensors.
-     *
-     * @param[in]  input  Source tensor. 3 lower dimensions represent a single input with dimensions [width, height, channels].
-     *                    Data types supported: QASYMM8/QASYMM8_SIGNED/F16/F32.
-     * @param[out] output Destination tensor. Data type supported: same as @p input
-     * @param[in]  mean   Mean values tensor. 1 dimension with size equal to the number of input channels. Data types supported: same as @p input
-     * @param[in]  std    Standard deviation values tensor. 1 dimension with size equal to the number of input channels.
-     *                    Data types supported: same as @p input
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, const ICLTensor *mean, const ICLTensor *std);
-    /** Set the input and output tensors.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Source tensor. 3 lower dimensions represent a single input with dimensions [width, height, channels].
-     *                             Data types supported: QASYMM8/QASYMM8_SIGNED/F16/F32.
-     * @param[out] output          Destination tensor. Data type supported: same as @p input
-     * @param[in]  mean            Mean values tensor. 1 dimension with size equal to the number of input channels. Data types supported: same as @p input
-     * @param[in]  std             Standard deviation values tensor. 1 dimension with size equal to the number of input channels.
-     *                             Data types supported: same as @p input
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, const ICLTensor *mean, const ICLTensor *std);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLNormalizePlanarYUVLayerKernel
-     *
-     * @param[in]  input  Source tensor info. 3 lower dimensions represent a single input with dimensions [width, height, channels].
-     *                    Data types supported: QASYMM8/QASYMM8_SIGNED/F16/F32.
-     * @param[out] output Destination tensor info. Data type supported: same as @p input
-     * @param[in]  mean   Mean values tensor info. 1 dimension with size equal to the number of input channels. Data types supported: same as @p input
-     * @param[in]  std    Standard deviation values tensor info. 1 dimension with size equal to the number of input channels.
-     *                    Data types supported: same as @p input
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output, const ITensorInfo *mean, const ITensorInfo *std);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;
-    ICLTensor       *_output;
-    const ICLTensor *_mean;
-    const ICLTensor *_std;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLNORMALIZEPLANARYUVLAYERKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLPadLayerKernel.h b/arm_compute/core/CL/kernels/CLPadLayerKernel.h
deleted file mode 100644
index 5bf5841803..0000000000
--- a/arm_compute/core/CL/kernels/CLPadLayerKernel.h
+++ /dev/null
@@ -1,96 +0,0 @@
-/*
- * Copyright (c) 2019-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLPADLAYERKERNEL_H
-#define ARM_COMPUTE_CLPADLAYERKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the PadLayer function. */
-class CLPadLayerKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLPadLayerKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLPadLayerKernel(const CLPadLayerKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLPadLayerKernel &operator=(const CLPadLayerKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLPadLayerKernel(CLPadLayerKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLPadLayerKernel &operator=(CLPadLayerKernel &&) = default;
-    /** Default destructor */
-    ~CLPadLayerKernel() = default;
-    /** Set the input and output tensor.
-     *
-     * @param[in]  input          Source tensor. Data types supported: All.
-     * @param[out] output         Output tensor. Data type supported: same as @p input
-     * @param[in]  padding        The padding for each spatial dimension of the input tensor. The pair padding[i]
-     *                            specifies the front and the end padding in the i-th dimension.
-     * @param[in]  constant_value (Optional) Constant value to be used for the padding.
-     * @param[in]  mode           (Optional) Controls whether the padding should be filled with @p constant_value using CONSTANT,
-     *                            or reflect the input, either including the border values (SYMMETRIC) or not (REFLECT).
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, const PaddingList &padding, PixelValue constant_value = PixelValue(), PaddingMode mode = PaddingMode::CONSTANT);
-    /** Set the input and output tensor.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Source tensor. Data types supported: All.
-     * @param[out] output          Output tensor. Data type supported: same as @p input
-     * @param[in]  padding         The padding for each spatial dimension of the input tensor. The pair padding[i]
-     *                             specifies the front and the end padding in the i-th dimension.
-     * @param[in]  constant_value  (Optional) Constant value to be used for the padding.
-     * @param[in]  mode            (Optional) Controls whether the padding should be filled with @p constant_value using CONSTANT,
-     *                             or reflect the input, either including the border values (SYMMETRIC) or not (REFLECT).
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, const PaddingList &padding, PixelValue constant_value = PixelValue(),
-                   PaddingMode mode = PaddingMode::CONSTANT);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLPadLayerKernel
-     *
-     * @param[in] input          Source tensor info. Data types supported: All.
-     * @param[in] output         Output tensor info. Data type supported: same as @p input
-     * @param[in] padding        The padding for each spatial dimension of the input tensor. The pair padding[i]
-     *                           specifies the front and the end padding in the i-th dimension.
-     * @param[in] constant_value (Optional) Constant value to be used for the padding.
-     * @param[in] mode           (Optional) Controls whether the padding should be filled with @p constant_value using CONSTANT,
-     *                            or reflect the input, either including the border values (SYMMETRIC) or not (REFLECT).
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output, const PaddingList &padding, PixelValue constant_value = PixelValue(), PaddingMode mode = PaddingMode::CONSTANT);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;
-    ICLTensor       *_output;
-    int              _input_start_x;
-    int              _input_start_y;
-    bool             _4d_enabled;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLPADLAYERKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLPermuteKernel.h b/arm_compute/core/CL/kernels/CLPermuteKernel.h
deleted file mode 100644
index bb841b1c83..0000000000
--- a/arm_compute/core/CL/kernels/CLPermuteKernel.h
+++ /dev/null
@@ -1,90 +0,0 @@
-/*
- * Copyright (c) 2018-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLPERMUTEKERNEL_H
-#define ARM_COMPUTE_CLPERMUTEKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** OpenCL kernel to perform tensor permutation.
- *
- * Permutes given a permutation vector
- */
-class CLPermuteKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLPermuteKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLPermuteKernel(const CLPermuteKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLPermuteKernel &operator=(const CLPermuteKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLPermuteKernel(CLPermuteKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLPermuteKernel &operator=(CLPermuteKernel &&) = default;
-    /** Set the input and output of the kernel.
-     *
-     * @note Arbitrary permutation vectors are supported with rank not greater than 4
-     *
-     * @param[in] input  The input tensor to permute. Data types supported: All.
-     * @param[in] output The output tensor. Data types supported: Same as @p input
-     * @param[in] perm   Permutation vector
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, const PermutationVector &perm);
-    /** Set the input and output of the kernel.
-     *
-     * @note Arbitrary permutation vectors are supported with rank not greater than 4
-     *
-     * @param[in] compile_context The compile context to be used.
-     * @param[in] input           The input tensor to permute. Data types supported: All.
-     * @param[in] output          The output tensor. Data types supported: Same as @p input
-     * @param[in] perm            Permutation vector
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, const PermutationVector &perm);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLPermuteKernel
-     *
-     * @note Arbitrary permutation vectors are supported with rank not greater than 4
-     *
-     * @param[in] input  First tensor input info. Data types supported: All.
-     * @param[in] output Output tensor info. Data types supported: same as @p input.
-     * @param[in] perm   Permutation vector
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output, const PermutationVector &perm);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;
-    ICLTensor        *_output;
-    PermutationVector _perm;
-};
-} // arm_compute
-#endif /*ARM_COMPUTE_CLPERMUTEKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLPixelWiseMultiplicationKernel.h b/arm_compute/core/CL/kernels/CLPixelWiseMultiplicationKernel.h
deleted file mode 100644
index 6b5bd11bde..0000000000
--- a/arm_compute/core/CL/kernels/CLPixelWiseMultiplicationKernel.h
+++ /dev/null
@@ -1,196 +0,0 @@
-/*
- * Copyright (c) 2016-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLPIXELWISEMULTIPLICATIONKERNEL_H
-#define ARM_COMPUTE_CLPIXELWISEMULTIPLICATIONKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-#include "arm_compute/core/Types.h"
-
-namespace arm_compute
-{
-// Forward declarations
-class ICLTensor;
-
-/** Interface for the pixelwise multiplication kernel. */
-class CLPixelWiseMultiplicationKernel : public ICLKernel
-{
-public:
-    /** Default constructor.*/
-    CLPixelWiseMultiplicationKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLPixelWiseMultiplicationKernel(const CLPixelWiseMultiplicationKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLPixelWiseMultiplicationKernel &operator=(const CLPixelWiseMultiplicationKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLPixelWiseMultiplicationKernel(CLPixelWiseMultiplicationKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLPixelWiseMultiplicationKernel &operator=(CLPixelWiseMultiplicationKernel &&) = default;
-    /** Initialise the kernel's input, output and border mode.
-     *
-     * Valid configurations (Input1,Input2) -> Output :
-     *
-     *   - (U8,U8)                         -> U8
-     *   - (U8,U8)                         -> S16
-     *   - (U8,S16)                        -> S16
-     *   - (S16,U8)                        -> S16
-     *   - (S16,S16)                       -> S16
-     *   - (F16,F16)                       -> F16
-     *   - (F32,F32)                       -> F32
-     *   - (QASYMM8,QASYMM8)               -> QASYMM8
-     *   - (QASYMM8_SIGNED,QASYMM8_SIGNED) -> QASYMM8_SIGNED
-     *   - (QSYMM16,QSYMM16)               -> QSYMM16
-     *   - (QSYMM16,QSYMM16)               -> S32
-     *
-     * @param[in]  input1          An input tensor info. Data types supported: U8/QASYMM8/QASYMM8_SIGNED/S16/QSYMM16/F16/F32.
-     * @param[in]  input2          An input tensor info. Data types supported: U8/QASYMM8/QASYMM8_SIGNED/S16/QSYMM16/F16/F32.
-     * @param[out] output          The output tensor info. Data types supported: U8/QASYMM8/QASYMM8_SIGNED/S16/QSYMM16/F16/F32.
-     * @param[in]  scale           Scale to apply after multiplication.
-     *                             Scale must be positive and its value must be either 1/255 or 1/2^n where n is between 0 and 15.
-     * @param[in]  overflow_policy Overflow policy. Supported overflow policies: Wrap, Saturate
-     * @param[in]  rounding_policy Rounding policy. Supported rounding modes: to zero, to nearest even.
-     * @param[in]  act_info        (Optional) Activation layer information in case of a fused activation.
-     */
-    void configure(ITensorInfo *input1, ITensorInfo *input2, ITensorInfo *output, float scale,
-                   ConvertPolicy overflow_policy, RoundingPolicy rounding_policy, const ActivationLayerInfo &act_info = ActivationLayerInfo());
-    /** Initialise the kernel's input, output and border mode.
-     *
-     * Valid configurations (Input1,Input2) -> Output :
-     *
-     *   - (U8,U8)                         -> U8
-     *   - (U8,U8)                         -> S16
-     *   - (U8,S16)                        -> S16
-     *   - (S16,U8)                        -> S16
-     *   - (S16,S16)                       -> S16
-     *   - (F16,F16)                       -> F16
-     *   - (F32,F32)                       -> F32
-     *   - (QASYMM8,QASYMM8)               -> QASYMM8
-     *   - (QASYMM8_SIGNED,QASYMM8_SIGNED) -> QASYMM8_SIGNED
-     *   - (QSYMM16,QSYMM16)               -> QSYMM16
-     *   - (QSYMM16,QSYMM16)               -> S32
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input1          An input tensor info. Data types supported: U8/QASYMM8/QASYMM8_SIGNED/S16/QSYMM16/F16/F32.
-     * @param[in]  input2          An input tensor info. Data types supported: U8/QASYMM8/QASYMM8_SIGNED/S16/QSYMM16/F16/F32.
-     * @param[out] output          The output tensor info. Data types supported: U8/QASYMM8/QASYMM8_SIGNED/S16/QSYMM16/F16/F32.
-     * @param[in]  scale           Scale to apply after multiplication.
-     *                             Scale must be positive and its value must be either 1/255 or 1/2^n where n is between 0 and 15.
-     * @param[in]  overflow_policy Overflow policy. Supported overflow policies: Wrap, Saturate
-     * @param[in]  rounding_policy Rounding policy. Supported rounding modes: to zero, to nearest even.
-     * @param[in]  act_info        (Optional) Activation layer information in case of a fused activation.
-     */
-    void configure(const CLCompileContext &compile_context, ITensorInfo *input1, ITensorInfo *input2, ITensorInfo *output, float scale,
-                   ConvertPolicy overflow_policy, RoundingPolicy rounding_policy, const ActivationLayerInfo &act_info = ActivationLayerInfo());
-    /** Static function to check if given info will lead to a valid configuration of @ref CLPixelWiseMultiplicationKernel
-     *
-     * Valid configurations (Input1,Input2) -> Output :
-     *
-     *   - (U8,U8)                         -> U8
-     *   - (U8,U8)                         -> S16
-     *   - (U8,S16)                        -> S16
-     *   - (S16,U8)                        -> S16
-     *   - (S16,S16)                       -> S16
-     *   - (F16,F16)                       -> F16
-     *   - (F32,F32)                       -> F32
-     *   - (QASYMM8,QASYMM8)               -> QASYMM8
-     *   - (QASYMM8_SIGNED,QASYMM8_SIGNED) -> QASYMM8_SIGNED
-     *   - (QSYMM16,QSYMM16)               -> QSYMM16
-     *   - (QSYMM16,QSYMM16)               -> S32
-     *
-     * @param[in] input1          An input tensor info. Data types supported: U8/QASYMM8/QASYMM8_SIGNED/S16/QSYMM16/F16/F32.
-     * @param[in] input2          An input tensor info. Data types supported: U8/QASYMM8/QASYMM8_SIGNED/S16/QSYMM16/F16/F32.
-     * @param[in] output          The output tensor info. Data types supported: U8/QASYMM8/QASYMM8_SIGNED/S16/QSYMM16/F16/F32.
-     * @param[in] scale           Scale to apply after multiplication.
-     *                            Scale must be positive and its value must be either 1/255 or 1/2^n where n is between 0 and 15.
-     * @param[in] overflow_policy Overflow policy. Supported overflow policies: Wrap, Saturate
-     * @param[in] rounding_policy Rounding policy. Supported rounding modes: to zero, to nearest even.
-     * @param[in] act_info        (Optional) Activation layer information in case of a fused activation.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input1, const ITensorInfo *input2, const ITensorInfo *output, float scale,
-                           ConvertPolicy overflow_policy, RoundingPolicy rounding_policy, const ActivationLayerInfo &act_info = ActivationLayerInfo());
-
-    // Inherited methods overridden:
-    void run_op(ITensorPack &tensors, const Window &window, cl::CommandQueue &queue) override;
-    BorderSize border_size() const override;
-
-private:
-    const ITensorInfo *_input1;
-    const ITensorInfo *_input2;
-    ITensorInfo       *_output;
-};
-
-/** Interface for the complex pixelwise multiplication kernel. */
-class CLComplexPixelWiseMultiplicationKernel : public ICLKernel
-{
-public:
-    /** Default constructor.*/
-    CLComplexPixelWiseMultiplicationKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLComplexPixelWiseMultiplicationKernel(const CLComplexPixelWiseMultiplicationKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLComplexPixelWiseMultiplicationKernel &operator=(const CLComplexPixelWiseMultiplicationKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLComplexPixelWiseMultiplicationKernel(CLComplexPixelWiseMultiplicationKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLComplexPixelWiseMultiplicationKernel &operator=(CLComplexPixelWiseMultiplicationKernel &&) = default;
-    /** Initialise the kernel's input, output and border mode.
-     *
-     * @param[in]  input1   An input tensor info. Data types supported: F32. Number of channels supported: 2.
-     * @param[in]  input2   An input tensor info. Data types supported: same as @p input1. Number of channels supported: same as @p input1.
-     * @param[out] output   The output tensor info. Data types supported: same as @p input1. Number of channels supported: same as @p input1.
-     * @param[in]  act_info (Optional) Activation layer information in case of a fused activation.
-     */
-    void configure(ITensorInfo *input1, ITensorInfo *input2, ITensorInfo *output, const ActivationLayerInfo &act_info = ActivationLayerInfo());
-    /** Initialise the kernel's input, output and border mode.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input1          An input tensor info. Data types supported: F32. Number of channels supported: 2.
-     * @param[in]  input2          An input tensor info. Data types supported: same as @p input1. Number of channels supported: same as @p input1.
-     * @param[out] output          The output tensor info. Data types supported: same as @p input1. Number of channels supported: same as @p input1.
-     * @param[in]  act_info        (Optional) Activation layer information in case of a fused activation.
-     */
-    void configure(const CLCompileContext &compile_context, ITensorInfo *input1, ITensorInfo *input2, ITensorInfo *output, const ActivationLayerInfo &act_info = ActivationLayerInfo());
-    /** Static function to check if given info will lead to a valid configuration of @ref CLComplexPixelWiseMultiplicationKernel
-     *
-     * @param[in] input1   An input tensor info. Data types supported: F32. Number of channels supported: 2.
-     * @param[in] input2   An input tensor info. Data types supported: same as @p input1. Number of channels supported: same as @p input1.
-     * @param[in] output   The output tensor info. Data types supported: same as @p input1. Number of channels supported: same as @p input1.
-     * @param[in] act_info (Optional) Activation layer information in case of a fused activation.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input1, const ITensorInfo *input2, const ITensorInfo *output, const ActivationLayerInfo &act_info = ActivationLayerInfo());
-
-    // Inherited methods overridden:
-    void run_op(ITensorPack &tensors, const Window &window, cl::CommandQueue &queue) override;
-    BorderSize border_size() const override;
-
-private:
-    const ITensorInfo *_input1;
-    const ITensorInfo *_input2;
-    ITensorInfo       *_output;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLPIXELWISEMULTIPLICATIONKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLPoolingLayerKernel.h b/arm_compute/core/CL/kernels/CLPoolingLayerKernel.h
deleted file mode 100644
index 85585e4587..0000000000
--- a/arm_compute/core/CL/kernels/CLPoolingLayerKernel.h
+++ /dev/null
@@ -1,96 +0,0 @@
-/*
- * Copyright (c) 2017-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLPOOLINGLAYERKERNEL_H
-#define ARM_COMPUTE_CLPOOLINGLAYERKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-#include "arm_compute/core/Error.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the pooling layer kernel */
-class CLPoolingLayerKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLPoolingLayerKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLPoolingLayerKernel(const CLPoolingLayerKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLPoolingLayerKernel &operator=(const CLPoolingLayerKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLPoolingLayerKernel(CLPoolingLayerKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLPoolingLayerKernel &operator=(CLPoolingLayerKernel &&) = default;
-    /** Default destructor */
-    ~CLPoolingLayerKernel() = default;
-
-    /** Set the input and output tensors.
-     *
-     *
-     * @param[in]  input     Source tensor. Data types supported: QASYMM8/QASYMM8_SIGNED/F16/F32.
-     * @param[out] output    Destination tensor. Data types supported: Same as @p input.
-     * @param[in]  pool_info Contains pooling operation information described in @ref PoolingLayerInfo.
-     * @param[out] indices   (optional) The indices of the maximal values. Data type supported: U32.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, const PoolingLayerInfo &pool_info, ICLTensor *indices = nullptr);
-    /** Set the input and output tensors.
-     *
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Source tensor. Data types supported: QASYMM8/QASYMM8_SIGNED/F16/F32.
-     * @param[out] output          Destination tensor. Data types supported: Same as @p input.
-     * @param[in]  pool_info       Contains pooling operation information described in @ref PoolingLayerInfo.
-     * @param[out] indices         (optional) The indices of the maximal values. Data type supported: U32.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, const PoolingLayerInfo &pool_info, ICLTensor *indices = nullptr);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLPoolingLayerKernel
-     *
-     * @param[in] input     Source tensor info. Data types supported: QASYMM8/QASYMM8_SIGNED/F16/F32.
-     * @param[in] output    Destination tensor info. Data types supported: Same as @p input.
-     * @param[in] pool_info Contains pooling operation information described in @ref PoolingLayerInfo.
-     * @param[in] indices   (optional) The indices of the maximal values. Data type supported: U32.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output, const PoolingLayerInfo &pool_info, const ITensorInfo *indices = nullptr);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-    BorderSize border_size() const override;
-
-public:
-    const ICLTensor *_input;
-    ICLTensor       *_output;
-    ICLTensor       *_indices;
-    PoolingLayerInfo _pool_info;
-    DataLayout       _data_layout;
-    BorderSize       _border_size;
-    unsigned int     _num_elems_processed_per_iteration;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLPOOLINGLAYERKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLPriorBoxLayerKernel.h b/arm_compute/core/CL/kernels/CLPriorBoxLayerKernel.h
deleted file mode 100644
index b4a69ac496..0000000000
--- a/arm_compute/core/CL/kernels/CLPriorBoxLayerKernel.h
+++ /dev/null
@@ -1,99 +0,0 @@
-/*
- * Copyright (c) 2018-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLPRIORBOXLAYERKERNEL_H
-#define ARM_COMPUTE_CLPRIORBOXLAYERKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the PriorBox layer kernel. */
-class CLPriorBoxLayerKernel : public ICLKernel
-{
-public:
-    /** Constructor */
-    CLPriorBoxLayerKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLPriorBoxLayerKernel(const CLPriorBoxLayerKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLPriorBoxLayerKernel &operator=(const CLPriorBoxLayerKernel &) = delete;
-    /** Default Move Constructor. */
-    CLPriorBoxLayerKernel(CLPriorBoxLayerKernel &&) = default;
-    /** Default move assignment operator */
-    CLPriorBoxLayerKernel &operator=(CLPriorBoxLayerKernel &&) = default;
-    /** Default destructor */
-    ~CLPriorBoxLayerKernel() = default;
-
-    /** Set the input and output tensors.
-     *
-     * @param[in]  input1        First source tensor. Data types supported: F32. Data layouts supported: NCHW/NHWC.
-     * @param[in]  input2        Second source tensor. Data types and layouts supported: same as @p input1
-     * @param[out] output        Destination tensor. Output dimensions are [W * H * num_priors * 4, 2]. Data types and layouts supported: same as @p input1
-     * @param[in]  info          Prior box layer info.
-     * @param[in]  min           Minimum prior box values
-     * @param[in]  max           Maximum prior box values
-     * @param[in]  aspect_ratios Aspect ratio values
-     */
-    void configure(const ICLTensor *input1, const ICLTensor *input2, ICLTensor *output, const PriorBoxLayerInfo &info, cl::Buffer *min, cl::Buffer *max, cl::Buffer *aspect_ratios);
-    /** Set the input and output tensors.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input1          First source tensor. Data types supported: F32. Data layouts supported: NCHW/NHWC.
-     * @param[in]  input2          Second source tensor. Data types and layouts supported: same as @p input1
-     * @param[out] output          Destination tensor. Output dimensions are [W * H * num_priors * 4, 2]. Data types and layouts supported: same as @p input1
-     * @param[in]  info            Prior box layer info.
-     * @param[in]  min             Minimum prior box values
-     * @param[in]  max             Maximum prior box values
-     * @param[in]  aspect_ratios   Aspect ratio values
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input1, const ICLTensor *input2, ICLTensor *output, const PriorBoxLayerInfo &info, cl::Buffer *min, cl::Buffer *max,
-                   cl::Buffer *aspect_ratios);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLPriorBoxLayerKernel
-     *
-     * @param[in] input1 First source tensor info. Data types supported: F32. Data layouts supported: NCHW/NHWC.
-     * @param[in] input2 Second source tensor info. Data types and layouts supported: same as @p input1
-     * @param[in] output Destination tensor info. Output dimensions are [W * H * num_priors * 4, 2]. Data type supported: same as @p input1
-     * @param[in] info   Prior box layer info.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input1, const ITensorInfo *input2, const ITensorInfo *output, const PriorBoxLayerInfo &info);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input1;
-    const ICLTensor *_input2;
-    ICLTensor        *_output;
-    PriorBoxLayerInfo _info;
-    int               _num_priors;
-    cl::Buffer       *_min;
-    cl::Buffer       *_max;
-    cl::Buffer       *_aspect_ratios;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLPRIORBOXLAYERKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLQLSTMLayerNormalizationKernel.h b/arm_compute/core/CL/kernels/CLQLSTMLayerNormalizationKernel.h
deleted file mode 100644
index 51c50bc011..0000000000
--- a/arm_compute/core/CL/kernels/CLQLSTMLayerNormalizationKernel.h
+++ /dev/null
@@ -1,88 +0,0 @@
-/*
- * Copyright (c) 2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLQLSTMLAYERVNORMALIZATIONKERNEL_H
-#define ARM_COMPUTE_CLQLSTMLAYERVNORMALIZATIONKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the kernel to do layer normalization. */
-class CLQLSTMLayerNormalizationKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLQLSTMLayerNormalizationKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLQLSTMLayerNormalizationKernel(const CLQLSTMLayerNormalizationKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLQLSTMLayerNormalizationKernel &operator=(const CLQLSTMLayerNormalizationKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLQLSTMLayerNormalizationKernel(CLQLSTMLayerNormalizationKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLQLSTMLayerNormalizationKernel &operator=(CLQLSTMLayerNormalizationKernel &&) = default;
-    /** Default destructor */
-    ~CLQLSTMLayerNormalizationKernel() = default;
-    /** Initialise the kernel's input and outputs.
-     *
-     * @param[in]  input  Source tensor with 2 dimensions. Data types supported: QSYMM16.
-     * @param[out] output Destination tensor. Data type supported: same as @p input
-     * @param[in]  weight Weight tensor. Data types supported: Same as @p input.
-     * @param[in]  bias   Bias tensor. Data types supported: S32.
-     *
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, const ICLTensor *weight, const ICLTensor *bias);
-    /** Initialise the kernel's input and outputs.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Source tensor with 2 dimensions. Data types supported: QSYMM16.
-     * @param[out] output          Destination tensor. Data type supported: same as @p input
-     * @param[in]  weight          Weight tensor. Data types supported: Same as @p input.
-     * @param[in]  bias            Bias tensor. Data types supported: S32.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, const ICLTensor *weight, const ICLTensor *bias);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLQLSTMLayerNormalizationKernel
-     *
-     * @param[in] input  Source tensor info with 2 dimensions. Data types supported: QSYMM16.
-     * @param[in] output Destination info tensor. Data type supported: same as @p input
-     * @param[in] weight Weight info tensor. Data types supported: Same as @p input.
-     * @param[in] bias   Bias tensor info. Data types supported: S32.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output, const ITensorInfo *weight, const ITensorInfo *bias);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;
-    const ICLTensor *_weight;
-    const ICLTensor *_bias;
-    ICLTensor       *_output;
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLQLSTMLAYERVNORMALIZATIONKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLQuantizationLayerKernel.h b/arm_compute/core/CL/kernels/CLQuantizationLayerKernel.h
deleted file mode 100644
index b0144bf8b0..0000000000
--- a/arm_compute/core/CL/kernels/CLQuantizationLayerKernel.h
+++ /dev/null
@@ -1,86 +0,0 @@
-/*
- * Copyright (c) 2017-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLQUANTIZATIONLAYERKERNEL_H
-#define ARM_COMPUTE_CLQUANTIZATIONLAYERKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the quantization layer kernel.
- *
- * @note The implementation supports only 3D input tensors.
- */
-class CLQuantizationLayerKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLQuantizationLayerKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLQuantizationLayerKernel(const CLQuantizationLayerKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLQuantizationLayerKernel &operator=(const CLQuantizationLayerKernel &) = delete;
-    /** Default Move Constructor. */
-    CLQuantizationLayerKernel(CLQuantizationLayerKernel &&) = default;
-    /** Default move assignment operator */
-    CLQuantizationLayerKernel &operator=(CLQuantizationLayerKernel &&) = default;
-    /** Default destructor */
-    ~CLQuantizationLayerKernel() = default;
-    /** Set the input, output.
-     *
-     * @param[in]  input  Source tensor. Data types supported: QASYMM8/QASYMM8_SIGNED/F32/F16.
-     * @param[out] output Destination tensor with the same dimensions of input. Data types supported: QASYMM8/QASYMM8_SIGNED/QASYMM16.
-     *
-     * @note Output auto initialization is not supported by this kernel
-     */
-    void configure(const ICLTensor *input, ICLTensor *output);
-    /** Set the input, output.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Source tensor. Data types supported: QASYMM8/QASYMM8_SIGNED/F32/F16.
-     * @param[out] output          Destination tensor with the same dimensions of input. Data types supported: QASYMM8/QASYMM8_SIGNED/QASYMM16.
-     *
-     * @note Output auto initialization is not supported by this kernel
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLQuantizationLayerKernel
-     *
-     * @param[in] input  Input tensor info. Data types supported: QASYMM8/QASYMM8_SIGNED/F32/F16.
-     * @param[in] output Destination tensor info with the same dimensions of input. Data types supported: QASYMM8/QASYMM8_SIGNED/QASYMM16.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;
-    ICLTensor       *_output;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLQUANTIZATIONLAYERKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLROIAlignLayerKernel.h b/arm_compute/core/CL/kernels/CLROIAlignLayerKernel.h
deleted file mode 100644
index 6a0468d331..0000000000
--- a/arm_compute/core/CL/kernels/CLROIAlignLayerKernel.h
+++ /dev/null
@@ -1,110 +0,0 @@
-/*
- * Copyright (c) 2018-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLROIALIGNLAYERKERNEL_H
-#define ARM_COMPUTE_CLROIALIGNLAYERKERNEL_H
-
-#include "arm_compute/core/CL/ICLArray.h"
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the RoIAlign kernel.
- */
-class CLROIAlignLayerKernel : public ICLKernel
-{
-public:
-    /** Constructor */
-    CLROIAlignLayerKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLROIAlignLayerKernel(const CLROIAlignLayerKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLROIAlignLayerKernel &operator=(const CLROIAlignLayerKernel &) = delete;
-    /** Default Move Constructor. */
-    CLROIAlignLayerKernel(CLROIAlignLayerKernel &&) = default;
-    /** Default move assignment operator. */
-    CLROIAlignLayerKernel &operator=(CLROIAlignLayerKernel &&) = default;
-    /** Default destructor */
-    ~CLROIAlignLayerKernel() = default;
-
-    /** Set the input and output tensors.
-     *
-     * @param[in]  input     Source tensor. Data types supported: QASYMM8/QASYMM8_SIGNED/F16/F32.
-     * @param[in]  rois      ROIs tensor, it is a 2D tensor of size [5, N] (where N is the number of ROIs) containing top left and bottom right corner
-     *                       as coordinate of an image and batch_id of ROI [ batch_id, x1, y1, x2, y2 ].
-     *                       Data types supported: QASYMM16 with scale of 0.125 and 0 offset if @p input is QASYMM8/QASYMM8_SIGNED, otherwise same as @p input
-     * @param[out] output    Destination tensor. Data types supported: Same as @p input.
-     * @param[in]  pool_info Contains pooling operation information described in @ref ROIPoolingLayerInfo.
-     *
-     * @note The x and y dimensions of @p output tensor must be the same as @p pool_info 's pooled
-     * width and pooled height.
-     * @note The z dimensions of @p output tensor and @p input tensor must be the same.
-     * @note The fourth dimension of @p output tensor must be the same as the number of elements in @p rois array.
-     */
-    void configure(const ICLTensor *input, const ICLTensor *rois, ICLTensor *output, const ROIPoolingLayerInfo &pool_info);
-    /** Set the input and output tensors.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Source tensor. Data types supported: QASYMM8/QASYMM8_SIGNED/F16/F32.
-     * @param[in]  rois            ROIs tensor, it is a 2D tensor of size [5, N] (where N is the number of ROIs) containing top left and bottom right corner
-     *                             as coordinate of an image and batch_id of ROI [ batch_id, x1, y1, x2, y2 ].
-     *                             Data types supported: QASYMM16 with scale of 0.125 and 0 offset if @p input is QASYMM8/QASYMM8_SIGNED, otherwise same as @p input
-     * @param[out] output          Destination tensor. Data types supported: Same as @p input.
-     * @param[in]  pool_info       Contains pooling operation information described in @ref ROIPoolingLayerInfo.
-     *
-     * @note The x and y dimensions of @p output tensor must be the same as @p pool_info 's pooled
-     * width and pooled height.
-     * @note The z dimensions of @p output tensor and @p input tensor must be the same.
-     * @note The fourth dimension of @p output tensor must be the same as the number of elements in @p rois array.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, const ICLTensor *rois, ICLTensor *output, const ROIPoolingLayerInfo &pool_info);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLROIAlignLayerKernel
-     *
-     * @param[in] input     Source tensor info. Data types supported: QASYMM8/QASYMM8_SIGNED/F16/F32.
-     * @param[in] rois      ROIs tensor info. Data types supported: QASYMM16 with scale of 0.125 and 0 offset if @p input is QASYMM8/QASYMM8_SIGNED,
-     *                      otherwise same as @p input
-     * @param[in] output    Destination tensor info. Data types supported: Same as @p input.
-     * @param[in] pool_info Contains pooling operation information described in @ref ROIPoolingLayerInfo.
-     *
-     * @note The x and y dimensions of @p output tensor must be the same as @p pool_info 's pooled
-     * width and pooled height.
-     * @note The z dimensions of @p output tensor and @p input tensor must be the same.
-     * @note The fourth dimension of @p output tensor must be the same as the number of elements in @p rois array.
-     *
-     * @return a Status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *rois, ITensorInfo *output, const ROIPoolingLayerInfo &pool_info);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue);
-
-private:
-    const ICLTensor    *_input;
-    ICLTensor          *_output;
-    const ICLTensor    *_rois;
-    ROIPoolingLayerInfo _pool_info;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLROIALIGNLAYERKERNEL_H*/
diff --git a/arm_compute/core/CL/kernels/CLROIPoolingLayerKernel.h b/arm_compute/core/CL/kernels/CLROIPoolingLayerKernel.h
deleted file mode 100644
index ee422e10ee..0000000000
--- a/arm_compute/core/CL/kernels/CLROIPoolingLayerKernel.h
+++ /dev/null
@@ -1,92 +0,0 @@
-/*
- * Copyright (c) 2017-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLROIPOOLINGLAYERKERNEL_H
-#define ARM_COMPUTE_CLROIPOOLINGLAYERKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-#include "arm_compute/core/CL/ICLArray.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the ROI pooling layer kernel */
-class CLROIPoolingLayerKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLROIPoolingLayerKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLROIPoolingLayerKernel(const CLROIPoolingLayerKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLROIPoolingLayerKernel &operator=(const CLROIPoolingLayerKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLROIPoolingLayerKernel(CLROIPoolingLayerKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLROIPoolingLayerKernel &operator=(CLROIPoolingLayerKernel &&) = default;
-    /** Default destructor */
-    ~CLROIPoolingLayerKernel() = default;
-
-    /** Set the input and output tensors.
-     *
-     * @param[in]  input     Source tensor. Data types supported: F16/F32.
-     * @param[in]  rois      ROIs tensor, it is a 2D tensor of size [5, N] (where N is the number of ROIs) containing top left and bottom right corner
-     *                       as coordinate of an image and batch_id of ROI [ batch_id, x1, y1, x2, y2 ]. Data types supported: U16
-     * @param[out] output    Destination tensor. Data types supported: Same as @p input.
-     * @param[in]  pool_info Contains pooling operation information described in @ref ROIPoolingLayerInfo.
-     *
-     * @note The x and y dimensions of @p output tensor must be the same as @p pool_info 's pooled
-     * width and pooled height.
-     * @note The z dimensions of @p output tensor and @p input tensor must be the same.
-     * @note The fourth dimension of @p output tensor must be the same as the number of elements in @p rois array.
-     */
-    void configure(const ICLTensor *input, const ICLTensor *rois, ICLTensor *output, const ROIPoolingLayerInfo &pool_info);
-    /** Set the input and output tensors.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Source tensor. Data types supported: F16/F32.
-     * @param[in]  rois            ROIs tensor, it is a 2D tensor of size [5, N] (where N is the number of ROIs) containing top left and bottom right corner
-     *                             as coordinate of an image and batch_id of ROI [ batch_id, x1, y1, x2, y2 ]. Data types supported: U16
-     * @param[out] output          Destination tensor. Data types supported: Same as @p input.
-     * @param[in]  pool_info       Contains pooling operation information described in @ref ROIPoolingLayerInfo.
-     *
-     * @note The x and y dimensions of @p output tensor must be the same as @p pool_info 's pooled
-     * width and pooled height.
-     * @note The z dimensions of @p output tensor and @p input tensor must be the same.
-     * @note The fourth dimension of @p output tensor must be the same as the number of elements in @p rois array.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, const ICLTensor *rois, ICLTensor *output, const ROIPoolingLayerInfo &pool_info);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor    *_input;
-    const ICLTensor    *_rois;
-    ICLTensor          *_output;
-    ROIPoolingLayerInfo _pool_info;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLROIPOOLINGLAYERKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLRangeKernel.h b/arm_compute/core/CL/kernels/CLRangeKernel.h
deleted file mode 100644
index b5c64b2480..0000000000
--- a/arm_compute/core/CL/kernels/CLRangeKernel.h
+++ /dev/null
@@ -1,92 +0,0 @@
-/*
- * Copyright (c) 2018-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLRANGEKERNEL_H
-#define ARM_COMPUTE_CLRANGEKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-#include "arm_compute/core/Types.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Kernel class for Range
- *
- * range generates a 1-D tensor containing a sequence of numbers that begins at 'start' and extends by increments
- * of 'step' up to but not including 'end'.
- */
-class CLRangeKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLRangeKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLRangeKernel(const CLRangeKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLRangeKernel &operator=(const CLRangeKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLRangeKernel(CLRangeKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLRangeKernel &operator=(CLRangeKernel &&) = default;
-    /** Default destructor */
-    ~CLRangeKernel() = default;
-    /** Initialize the kernel's output tensor, start, end and step of the sequence.
-     *
-     * @param[out] output Output tensor. Data types supported: U8/S8/QASYMM8/U16/S16/U32/S32/F16/F32.
-     * @param[in]  start  The starting value of the sequence.
-     * @param[in]  end    The ending (not including) value of the sequence.
-     * @param[in]  step   The gap between each pair of values in the sequence.
-     */
-    void configure(ICLTensor *output, float start, float end, float step);
-    /** Initialize the kernel's output tensor, start, end and step of the sequence.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[out] output          Output tensor. Data types supported: U8/S8/QASYMM8/U16/S16/U32/S32/F16/F32.
-     * @param[in]  start           The starting value of the sequence.
-     * @param[in]  end             The ending (not including) value of the sequence.
-     * @param[in]  step            The gap between each pair of values in the sequence.
-     */
-    void configure(const CLCompileContext &compile_context, ICLTensor *output, float start, float end, float step);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLRangeKernel
-     *
-     * @param[in] output Output tensor info. Data types supported: U8/S8/QASYMM8/U16/S16/U32/S32/F16/F32.
-     * @param[in] start  The starting value of the sequence.
-     * @param[in] end    The ending (not including) value of the sequence.
-     * @param[in] step   The gap between each pair of values in the sequence.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *output, float start, float end, float step);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    float      _start;  /**< Start of sequence */
-    float      _end;    /**< End of sequence */
-    float      _step;   /**< Increment/step value */
-    ICLTensor *_output; /**< Destination tensor */
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLRANGEKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLReductionOperationKernel.h b/arm_compute/core/CL/kernels/CLReductionOperationKernel.h
deleted file mode 100644
index 2ecd1c9fd4..0000000000
--- a/arm_compute/core/CL/kernels/CLReductionOperationKernel.h
+++ /dev/null
@@ -1,99 +0,0 @@
-/*
- * Copyright (c) 2017-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLREDUCTIONOPERATIONKERNEL_H
-#define ARM_COMPUTE_CLREDUCTIONOPERATIONKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-#include "arm_compute/core/Types.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the reduction operation kernel
- */
-class CLReductionOperationKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLReductionOperationKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLReductionOperationKernel(const CLReductionOperationKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLReductionOperationKernel &operator=(const CLReductionOperationKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLReductionOperationKernel(CLReductionOperationKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLReductionOperationKernel &operator=(CLReductionOperationKernel &&) = default;
-    /** Default destructor */
-    ~CLReductionOperationKernel() = default;
-
-    /** Set the input and output tensors.
-     *
-     * @param[in]  input  Source tensor. Data types supported: QASYMM8/QASYMM8_SIGNED/S32/F16/F32.
-     * @param[out] output Destination tensor. Data types and data layouts supported: Same as @p input.
-     *                    Output will have the same number of dimensions as input.
-     * @param[in]  axis   Axis along which to reduce. Supported reduction axis : 0,1,2,3
-     * @param[in]  op     Reduction operation to perform. Operations supported: MEAN_SUM, PROD, SUM_SQUARE, SUM, MIN, MAX
-     * @param[in]  width  (Optional)  In case of x-axis we also need to provide the width of the input image.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, unsigned int axis, ReductionOperation op, unsigned int width = 0);
-    /** Set the input and output tensors.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Source tensor. Data types supported: QASYMM8/QASYMM8_SIGNED/S32/F16/F32.
-     * @param[out] output          Destination tensor. Data types and data layouts supported: Same as @p input.
-     *                             Output will have the same number of dimensions as input.
-     * @param[in]  axis            Axis along which to reduce. Supported reduction axis : 0,1,2,3
-     * @param[in]  op              Reduction operation to perform. Operations supported: MEAN_SUM, PROD, SUM_SQUARE, SUM, MIN, MAX
-     * @param[in]  width           (Optional)  In case of x-axis we also need to provide the width of the input image.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, unsigned int axis, ReductionOperation op, unsigned int width = 0);
-
-    /** Static function to check if given info will lead to a valid configuration of @ref CLReductionOperationKernel.
-     *
-     * @param[in] input  Source tensor info. Data types supported: QASYMM8/QASYMM8_SIGNED/S32/F16/F32.
-     * @param[in] output Destination tensor info. Data types and data layouts supported: Same as @p input.
-     *                   Output will have the same number of dimensions as input.
-     * @param[in] axis   Axis along which to reduce. Supported reduction axis : 0,1,2,3
-     * @param[in] op     Reduction operation to perform. Operations supported: MEAN_SUM, PROD, SUM_SQUARE, SUM, MIN, MAX
-     * @param[in] width  (Optional)  In case of x-axis we also need to provide the width of the input image.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output, unsigned int axis, ReductionOperation op, unsigned int width = 0);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-    BorderSize border_size() const override;
-
-private:
-    const ICLTensor   *_input;
-    ICLTensor         *_output;
-    unsigned int       _reduction_axis;
-    ReductionOperation _op;
-    BorderSize         _border_size;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLREDUCTIONOPERATIONKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLRemapKernel.h b/arm_compute/core/CL/kernels/CLRemapKernel.h
deleted file mode 100644
index fd261cd465..0000000000
--- a/arm_compute/core/CL/kernels/CLRemapKernel.h
+++ /dev/null
@@ -1,81 +0,0 @@
-/*
- * Copyright (c) 2017-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLREMAPKERNEL_H
-#define ARM_COMPUTE_CLREMAPKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-#include "arm_compute/core/Types.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** OpenCL kernel to perform a remap on a tensor */
-class CLRemapKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLRemapKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLRemapKernel(const CLRemapKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLRemapKernel &operator=(const CLRemapKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLRemapKernel(CLRemapKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLRemapKernel &operator=(CLRemapKernel &&) = default;
-    /** Initialize the kernel's input, output and border mode.
-     *
-     * @param[in]  input            Source tensor. Data types supported: U8.
-     * @param[in]  map_x            Map for X coordinates. Data types supported: F32.
-     * @param[in]  map_y            Map for Y coordinates. Data types supported: F32.
-     * @param[out] output           Destination tensor. Data types supported: U8. All but the lowest two dimensions must be the same size as in the input tensor, i.e. remapping is only performed within the XY-plane.
-     * @param[in]  policy           The interpolation type.
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const ICLTensor *input, const ICLTensor *map_x, const ICLTensor *map_y, ICLTensor *output, InterpolationPolicy policy, bool border_undefined);
-    /** Initialize the kernel's input, output and border mode.
-     *
-     * @param[in]  compile_context  The compile context to be used.
-     * @param[in]  input            Source tensor. Data types supported: U8.
-     * @param[in]  map_x            Map for X coordinates. Data types supported: F32.
-     * @param[in]  map_y            Map for Y coordinates. Data types supported: F32.
-     * @param[out] output           Destination tensor. Data types supported: U8. All but the lowest two dimensions must be the same size as in the input tensor, i.e. remapping is only performed within the XY-plane.
-     * @param[in]  policy           The interpolation type.
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, const ICLTensor *map_x, const ICLTensor *map_y, ICLTensor *output, InterpolationPolicy policy, bool border_undefined);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-    BorderSize border_size() const override;
-
-private:
-    const ICLTensor *_input;
-    ICLTensor       *_output;
-    const ICLTensor *_map_x;
-    const ICLTensor *_map_y;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLREMAPKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLReorgLayerKernel.h b/arm_compute/core/CL/kernels/CLReorgLayerKernel.h
deleted file mode 100644
index e3edc9f724..0000000000
--- a/arm_compute/core/CL/kernels/CLReorgLayerKernel.h
+++ /dev/null
@@ -1,90 +0,0 @@
-/*
- * Copyright (c) 2018-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLREORGLAYERKERNEL_H
-#define ARM_COMPUTE_CLREORGLAYERKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-#include "arm_compute/core/Types.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** OpenCL kernel to perform a reorg layer */
-class CLReorgLayerKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLReorgLayerKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers). */
-    CLReorgLayerKernel(const CLReorgLayerKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers). */
-    CLReorgLayerKernel &operator=(const CLReorgLayerKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLReorgLayerKernel(CLReorgLayerKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLReorgLayerKernel &operator=(CLReorgLayerKernel &&) = default;
-    /** Initialize the kernel's input, output.
-     *
-     * @param[in]  input  Source tensor. Data types supported: All.
-     * @param[out] output Destination tensor with tensor shape:
-     *                    [width_input / stride, height_input / stride, channels_input * stride * stride, batch_size]. This means the output has
-     *                    the same number of input elements. Data types supported: same as @p input.
-     * @param[in]  stride Stride value to use for reorganizing the values in the output tensor.
-     *                    It defines the spatial distance between 2 consecutive pixels in the x and y direction
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, int32_t stride);
-    /** Initialize the kernel's input, output.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Source tensor. Data types supported: All.
-     * @param[out] output          Destination tensor with tensor shape:
-     *                             [width_input / stride, height_input / stride, channels_input * stride * stride, batch_size]. This means the output has
-     *                             the same number of input elements. Data types supported: same as @p input.
-     * @param[in]  stride          Stride value to use for reorganizing the values in the output tensor.
-     *                             It defines the spatial distance between 2 consecutive pixels in the x and y direction
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, int32_t stride);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLReorgLayerKernel
-     *
-     * @param[in] input  Source tensor. Data types supported: All.
-     * @param[in] output Destination tensor with tensor shape:
-     *                   [width_input / stride, height_input / stride, channels_input * stride * stride, batch_size]. This means the output has
-     *                   the same number of input elements. Data types supported: same as @p input. Data types supported: same as @p input.
-     * @param[in] stride Stride value to use for reorganizing the values in the output tensor
-     *                   It defines the spatial distance between 2 consecutive pixels in the x and y direction
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output, int32_t stride);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;
-    ICLTensor       *_output;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLREORGLAYERKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLReshapeLayerKernel.h b/arm_compute/core/CL/kernels/CLReshapeLayerKernel.h
deleted file mode 100644
index 6e3f255c52..0000000000
--- a/arm_compute/core/CL/kernels/CLReshapeLayerKernel.h
+++ /dev/null
@@ -1,59 +0,0 @@
-/*
- * Copyright (c) 2017-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLRESHAPELAYERKERNEL_H
-#define ARM_COMPUTE_CLRESHAPELAYERKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-#include "arm_compute/core/Types.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the kernel to perform tensor reshaping */
-class CLReshapeLayerKernel : public ICLKernel
-{
-public:
-    /** Set the input and output of the kernel
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Source tensor info. Data type supported: All.
-     * @param[out] output          Destination tensor info. Data type supported: Same as @p input
-     */
-    void configure(const CLCompileContext &compile_context, const ITensorInfo *input, ITensorInfo *output);
-
-    /** Static function to check if given info will lead to a valid configuration of @ref CLReshapeLayerKernel
-     *
-     * @param[in] input  Source tensor info. Data type supported: All
-     * @param[in] output Destination tensor info. Data type supported: Same as @p input
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output);
-
-    // Inherited methods overridden:
-    void run_op(ITensorPack &tensors, const Window &window, cl::CommandQueue &queue) override;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLRESHAPELAYERKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLReverseKernel.h b/arm_compute/core/CL/kernels/CLReverseKernel.h
deleted file mode 100644
index 17f1a4a20f..0000000000
--- a/arm_compute/core/CL/kernels/CLReverseKernel.h
+++ /dev/null
@@ -1,84 +0,0 @@
-/*
- * Copyright (c) 2018-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLREVERSEKERNEL_H
-#define ARM_COMPUTE_CLREVERSEKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the reverse kernel */
-class CLReverseKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLReverseKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLReverseKernel(const CLReverseKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLReverseKernel &operator=(const CLReverseKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLReverseKernel(CLReverseKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLReverseKernel &operator=(CLReverseKernel &&) = default;
-    /** Default destructor */
-    ~CLReverseKernel() = default;
-    /** Initialise the kernel's inputis and output
-     *
-     * @param[in]  input  Input tensor. Data types supported: All.
-     * @param[out] output Output tensor. Data type supported: Same as @p input
-     * @param[in]  axis   Axis tensor. Contains the indices of the dimensions to reverse. Data type supported: U32
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, const ICLTensor *axis);
-    /** Initialise the kernel's inputis and output
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Input tensor. Data types supported: All.
-     * @param[out] output          Output tensor. Data type supported: Same as @p input
-     * @param[in]  axis            Axis tensor. Contains the indices of the dimensions to reverse. Data type supported: U32
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, const ICLTensor *axis);
-
-    /** Static function to check if given info will lead to a valid configuration of @ref CLReverseKernel
-     *
-     * @param[in] input  Input tensor info. Data types supported: All.
-     * @param[in] output Output tensor info. Data type supported: Same as @p input
-     * @param[in] axis   Axis tensor info. Contains the indices of the dimensions to reverse. Data type supported: U32
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output, const ITensorInfo *axis);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-public:
-    const ICLTensor *_input;
-    ICLTensor       *_output;
-    const ICLTensor *_axis;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLREVERSEKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLScaleKernel.h b/arm_compute/core/CL/kernels/CLScaleKernel.h
deleted file mode 100644
index 79f7ed181a..0000000000
--- a/arm_compute/core/CL/kernels/CLScaleKernel.h
+++ /dev/null
@@ -1,93 +0,0 @@
-/*
- * Copyright (c) 2016-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLSCALEKERNEL_H
-#define ARM_COMPUTE_CLSCALEKERNEL_H
-
-#include "arm_compute/core/CL/ICLSimple2DKernel.h"
-#include "arm_compute/core/KernelDescriptors.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the scale kernel */
-class CLScaleKernel : public ICLSimple2DKernel
-{
-public:
-    /** Initialise the kernel's inputs, output and interpolation policy
-     *
-     * @param[in]  input  Source tensor. Data types supported: U8/QASYMM8/QASYMM8_SIGNED/S16/F16/F32
-     * @param[out] output Destination tensor. Data types supported: Same as @p input
-     *                    All but the lowest two dimensions must be the same size as in the input tensor, i.e. scaling is only performed within the XY-plane.
-     * @param[in]  info   @ref ScaleKernelInfo Kernel descriptor to be used to configure.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, const ScaleKernelInfo &info);
-    /** Initialise the kernel's inputs, output and interpolation policy
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Source tensor. Data types supported: U8/QASYMM8/QASYMM8_SIGNED/S16/F16/F32
-     * @param[out] output          Destination tensor. Data types supported: Same as @p input
-     *                             All but the lowest two dimensions must be the same size as in the input tensor, i.e. scaling is only performed within the XY-plane.
-     * @param[in]  info            @ref ScaleKernelInfo Kernel descriptor to be used to configure.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, const ScaleKernelInfo &info);
-
-    /** Static function to check if given info will lead to a valid configuration of @ref CLScaleKernel
-     *
-     * @param[in] input  Source tensor info. Data types supported: U8/QASYMM8/QASYMM8_SIGNED/S16/F16/F32
-     * @param[in] output Destination tensor info. Data types supported: Same as @p input
-     *                   All but the lowest two dimensions must be the same size as in the input tensor, i.e. scaling is only performed within the XY-plane.
-     * @param[in] info   @ref ScaleKernelInfo Kernel descriptor to be used to validate
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output, const ScaleKernelInfo &info);
-    /** Input tensor accessor.
-     *
-     * @return Pointer to input tensor.
-     */
-    const ICLTensor *input() const;
-    /** Output tensor accessor.
-     *
-     * @return Pointer to output tensor.
-     */
-    const ICLTensor *output() const;
-
-    // Inherited methods overridden:
-    BorderSize border_size() const override;
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-    // Getter for interpolation policy
-    InterpolationPolicy get_interpolation_policy() const
-    {
-        return _interpolation_policy;
-    }
-
-private:
-    InterpolationPolicy _interpolation_policy = InterpolationPolicy::BILINEAR;
-    DataLayout          _data_layout          = DataLayout::UNKNOWN;
-    bool                _align_corners        = false;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLSCALEKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLScharr3x3Kernel.h b/arm_compute/core/CL/kernels/CLScharr3x3Kernel.h
deleted file mode 100644
index 1af56a764e..0000000000
--- a/arm_compute/core/CL/kernels/CLScharr3x3Kernel.h
+++ /dev/null
@@ -1,97 +0,0 @@
-/*
- * Copyright (c) 2016-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLSCHARR3X3KERNEL_H
-#define ARM_COMPUTE_CLSCHARR3X3KERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the kernel to run a 3x3 Scharr filter on a tensor.
- *
- * @f[
- *      \mathbf{G}_x=\begin{vmatrix}
- *      -3 & 0 & +3\\
- *      -10& 0 & +10\\
- *      -3 & 0 & +3
- *      \end{vmatrix}
- * @f]
- * @f[
- *      \mathbf{G}_y=\begin{vmatrix}
- *      -3 & -10 & -3\\
- *       0 & 0 & 0\\
- *      +3 & +10 & +3
- *      \end{vmatrix}
- * @f]
- */
-class CLScharr3x3Kernel : public ICLKernel
-{
-public:
-    /** Default constructor: initialize all the pointers to nullptr and parameters to zero. */
-    CLScharr3x3Kernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLScharr3x3Kernel(const CLScharr3x3Kernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLScharr3x3Kernel &operator=(const CLScharr3x3Kernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLScharr3x3Kernel(CLScharr3x3Kernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLScharr3x3Kernel &operator=(CLScharr3x3Kernel &&) = default;
-    /** Initialise the kernel's source, destination and border.
-     *
-     * @note At least one of output_x or output_y must be set.
-     *
-     * @param[in]  input            Source tensor. Data types supported: U8.
-     * @param[out] output_x         (Optional) Destination tensor for the X gradient, Data types supported: S16.
-     * @param[out] output_y         (Optional) Destination tensor for the Y gradient, Data types supported: S16.
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output_x, ICLTensor *output_y, bool border_undefined);
-    /** Initialise the kernel's source, destination and border.
-     *
-     * @note At least one of output_x or output_y must be set.
-     *
-     * @param[in]  compile_context  The compile context to be used.
-     * @param[in]  input            Source tensor. Data types supported: U8.
-     * @param[out] output_x         (Optional) Destination tensor for the X gradient, Data types supported: S16.
-     * @param[out] output_y         (Optional) Destination tensor for the Y gradient, Data types supported: S16.
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output_x, ICLTensor *output_y, bool border_undefined);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-    BorderSize border_size() const override;
-
-private:
-    bool             _run_scharr_x; /**< Do we need to run Scharr X ? */
-    bool             _run_scharr_y; /**< Do we need to run Scharr Y ? */
-    const ICLTensor *_input;        /**< Input image */
-    ICLTensor       *_output_x;     /**< Output image for scharr X */
-    ICLTensor       *_output_y;     /**< Output image for scharr Y */
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLSCHARR3X3KERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLSelectKernel.h b/arm_compute/core/CL/kernels/CLSelectKernel.h
deleted file mode 100644
index 4015a273ea..0000000000
--- a/arm_compute/core/CL/kernels/CLSelectKernel.h
+++ /dev/null
@@ -1,94 +0,0 @@
-/*
- * Copyright (c) 2018-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLSELECTKERNEL_H
-#define ARM_COMPUTE_CLSELECTKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-#include "arm_compute/core/Types.h"
-
-namespace arm_compute
-{
-// Forward declarations
-class ICLTensor;
-
-/** OpenCL interface for executing the select kernel
- *
- * Select is computed by:
- * @f[ output(i) = condition(i) ? x(i) : y(i) @f]
- **/
-class CLSelectKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLSelectKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLSelectKernel(const CLSelectKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLSelectKernel &operator=(const CLSelectKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLSelectKernel(CLSelectKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLSelectKernel &operator=(CLSelectKernel &&) = default;
-    /** Default destructor */
-    ~CLSelectKernel() = default;
-    /** Initialise the kernel's inputs and output.
-     *
-     * @param[in]  c      Condition input tensor. Data types supported: U8.
-     * @param[in]  x      First input tensor. Data types supported: All.
-     * @param[out] y      Second input tensor. Data types supported: Same as @p x
-     * @param[in]  output Output tensor. Data types supported: Same as @p x.
-     */
-    void configure(const ICLTensor *c, const ICLTensor *x, const ICLTensor *y, ICLTensor *output);
-    /** Initialise the kernel's inputs and output.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  c               Condition input tensor. Data types supported: U8.
-     * @param[in]  x               First input tensor. Data types supported: All.
-     * @param[out] y               Second input tensor. Data types supported: Same as @p x
-     * @param[in]  output          Output tensor. Data types supported: Same as @p x.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *c, const ICLTensor *x, const ICLTensor *y, ICLTensor *output);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLSelectKernel
-     *
-     * @param[in] c      Condition input tensor. Data types supported: U8.
-     * @param[in] x      First input tensor. Data types supported: All.
-     * @param[in] y      Second input tensor. Data types supported: Same as @p x
-     * @param[in] output Output tensor. Data types supported: Same as @p x.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *c, const ITensorInfo *x, const ITensorInfo *y, const ITensorInfo *output);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_c;             /**< Condition tensor */
-    const ICLTensor *_x;             /**< Source tensor 1 */
-    const ICLTensor *_y;             /**< Source tensor 2 */
-    ICLTensor       *_output;        /**< Destination tensor */
-    bool             _has_same_rank; /**< Flag that indicates if condition tensor and other inputs have the same rank */
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLWHEREKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLSobel3x3Kernel.h b/arm_compute/core/CL/kernels/CLSobel3x3Kernel.h
deleted file mode 100644
index e24767852e..0000000000
--- a/arm_compute/core/CL/kernels/CLSobel3x3Kernel.h
+++ /dev/null
@@ -1,83 +0,0 @@
-/*
- * Copyright (c) 2016-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLSOBEL3X3KERNEL_H
-#define ARM_COMPUTE_CLSOBEL3X3KERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the kernel to run a 3x3 Sobel filter on a tensor. */
-class CLSobel3x3Kernel : public ICLKernel
-{
-public:
-    /** Default constructor: initialize all the pointers to nullptr and parameters to zero. */
-    CLSobel3x3Kernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLSobel3x3Kernel(const CLSobel3x3Kernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLSobel3x3Kernel &operator=(const CLSobel3x3Kernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLSobel3x3Kernel(CLSobel3x3Kernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLSobel3x3Kernel &operator=(CLSobel3x3Kernel &&) = default;
-    /** Default destructor */
-    ~CLSobel3x3Kernel() = default;
-    /** Initialise the kernel's source, destination and border.
-     *
-     * @note At least one of output_x or output_y must be set.
-     *
-     * @param[in]  input            Source tensor. Data types supported: U8.
-     * @param[out] output_x         (Optional) Destination tensor for the X gradient, Data types supported: S16.
-     * @param[out] output_y         (Optional) Destination tensor for the Y gradient, Data types supported: S16.
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output_x, ICLTensor *output_y, bool border_undefined);
-    /** Initialise the kernel's source, destination and border.
-     *
-     * @note At least one of output_x or output_y must be set.
-     *
-     * @param[in]  compile_context  The compile context to be used.
-     * @param[in]  input            Source tensor. Data types supported: U8.
-     * @param[out] output_x         (Optional) Destination tensor for the X gradient, Data types supported: S16.
-     * @param[out] output_y         (Optional) Destination tensor for the Y gradient, Data types supported: S16.
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output_x, ICLTensor *output_y, bool border_undefined);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-    BorderSize border_size() const override;
-
-private:
-    const ICLTensor *_input;       /**< Input tensor */
-    ICLTensor       *_output_x;    /**< Output tensor for Sobel X */
-    ICLTensor       *_output_y;    /**< Output tensor for Sobel Y */
-    bool             _run_sobel_x; /**< Do we need to run Sobel X ? */
-    bool             _run_sobel_y; /**< Do we need to run Sobel Y ? */
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLSOBEL3X3KERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLSobel5x5Kernel.h b/arm_compute/core/CL/kernels/CLSobel5x5Kernel.h
deleted file mode 100644
index 82831ed14d..0000000000
--- a/arm_compute/core/CL/kernels/CLSobel5x5Kernel.h
+++ /dev/null
@@ -1,139 +0,0 @@
-/*
- * Copyright (c) 2016-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLSOBEL5X5KERNEL_H
-#define ARM_COMPUTE_CLSOBEL5X5KERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the kernel to run the horizontal pass of 5x5 Sobel filter on a tensor. */
-class CLSobel5x5HorKernel : public ICLKernel
-{
-public:
-    /** Default constructor: initialize all the pointers to nullptr and parameters to zero. */
-    CLSobel5x5HorKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLSobel5x5HorKernel(const CLSobel5x5HorKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLSobel5x5HorKernel &operator=(const CLSobel5x5HorKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLSobel5x5HorKernel(CLSobel5x5HorKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLSobel5x5HorKernel &operator=(CLSobel5x5HorKernel &&) = default;
-    /** Default destructor */
-    ~CLSobel5x5HorKernel() = default;
-
-    /** Initialise the kernel's source, destination and border.
-     *
-     * @note At least one of output_x or output_y must be set.
-     *
-     * @param[in]  input            Source tensor. Data types supported: U8.
-     * @param[out] output_x         (Optional) Destination tensor for the X gradient, Data types supported: S16.
-     * @param[out] output_y         (Optional) Destination tensor for the Y gradient, Data types supported: S16.
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output_x, ICLTensor *output_y, bool border_undefined);
-    /** Initialise the kernel's source, destination and border.
-     *
-     * @note At least one of output_x or output_y must be set.
-     *
-     * @param[in]  compile_context  The compile context to be used.
-     * @param[in]  input            Source tensor. Data types supported: U8.
-     * @param[out] output_x         (Optional) Destination tensor for the X gradient, Data types supported: S16.
-     * @param[out] output_y         (Optional) Destination tensor for the Y gradient, Data types supported: S16.
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output_x, ICLTensor *output_y, bool border_undefined);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-    BorderSize border_size() const override;
-
-private:
-    const ICLTensor *_input;       /**< Input tensor */
-    ICLTensor       *_output_x;    /**< X output of horizontal pass */
-    ICLTensor       *_output_y;    /**< Y output of horizontal pass */
-    bool             _run_sobel_x; /**< Do we need to run Sobel X ? */
-    bool             _run_sobel_y; /**< Do we need to run Sobel Y ? */
-    BorderSize       _border_size; /**< Border size */
-};
-
-/** Interface for the kernel to run the vertical pass of 5x5 Sobel filter on a tensor. */
-class CLSobel5x5VertKernel : public ICLKernel
-{
-public:
-    /** Default constructor: initialize all the pointers to nullptr and parameters to zero. */
-    CLSobel5x5VertKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLSobel5x5VertKernel(const CLSobel5x5VertKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLSobel5x5VertKernel &operator=(const CLSobel5x5VertKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLSobel5x5VertKernel(CLSobel5x5VertKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLSobel5x5VertKernel &operator=(CLSobel5x5VertKernel &&) = default;
-    /** Default destructor */
-    ~CLSobel5x5VertKernel() = default;
-
-    /** Initialise the kernel's source, destination and border.
-     *
-     * @note At least one of output_x or output_y must be set and the corresponding input.
-     *
-     * @param[in]  input_x          (Optional) Input for X (X output of horizontal pass). Data types supported: S16.
-     * @param[in]  input_y          (Optional) Input for Y (Y output of horizontal pass). Data types supported: S16.
-     * @param[out] output_x         (Optional) Destination tensor for the X gradient, Data types supported: S16.
-     * @param[out] output_y         (Optional) Destination tensor for the Y gradient, Data types supported: S16.
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const ICLTensor *input_x, const ICLTensor *input_y, ICLTensor *output_x, ICLTensor *output_y, bool border_undefined);
-    /** Initialise the kernel's source, destination and border.
-     *
-     * @note At least one of output_x or output_y must be set and the corresponding input.
-     *
-     * @param[in]  compile_context  The compile context to be used.
-     * @param[in]  input_x          (Optional) Input for X (X output of horizontal pass). Data types supported: S16.
-     * @param[in]  input_y          (Optional) Input for Y (Y output of horizontal pass). Data types supported: S16.
-     * @param[out] output_x         (Optional) Destination tensor for the X gradient, Data types supported: S16.
-     * @param[out] output_y         (Optional) Destination tensor for the Y gradient, Data types supported: S16.
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input_x, const ICLTensor *input_y, ICLTensor *output_x, ICLTensor *output_y, bool border_undefined);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-    BorderSize border_size() const override;
-
-private:
-    const ICLTensor *_input_x;     /**< X input (X output of the horizontal pass) */
-    const ICLTensor *_input_y;     /**< Y input (Y output of the horizontal pass) */
-    ICLTensor       *_output_x;    /**< X output of sobel */
-    ICLTensor       *_output_y;    /**< Y output of sobel */
-    bool             _run_sobel_x; /**< Do we need to run sobel X? */
-    bool             _run_sobel_y; /**< Do we need to run sobel Y? */
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLSOBEL5X5KERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLSobel7x7Kernel.h b/arm_compute/core/CL/kernels/CLSobel7x7Kernel.h
deleted file mode 100644
index d55993d1f0..0000000000
--- a/arm_compute/core/CL/kernels/CLSobel7x7Kernel.h
+++ /dev/null
@@ -1,139 +0,0 @@
-/*
- * Copyright (c) 2016-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLSOBEL7X7KERNEL_H
-#define ARM_COMPUTE_CLSOBEL7X7KERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the kernel to run the horizontal pass of 7x7 Sobel filter on a tensor. */
-class CLSobel7x7HorKernel : public ICLKernel
-{
-public:
-    /** Default constructor: initialize all the pointers to nullptr and parameters to zero. */
-    CLSobel7x7HorKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLSobel7x7HorKernel(const CLSobel7x7HorKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLSobel7x7HorKernel &operator=(const CLSobel7x7HorKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLSobel7x7HorKernel(CLSobel7x7HorKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLSobel7x7HorKernel &operator=(CLSobel7x7HorKernel &&) = default;
-    /** Default destructor */
-    ~CLSobel7x7HorKernel() = default;
-
-    /** Initialise the kernel's source, destination and border.
-     *
-     * @note At least one of output_x or output_y must be set.
-     *
-     * @param[in]  input            Source tensor. Data types supported: U8.
-     * @param[out] output_x         (Optional) Destination tensor for the X gradient, Data types supported: S32.
-     * @param[out] output_y         (Optional) Destination tensor for the Y gradient, Data types supported: S32.
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output_x, ICLTensor *output_y, bool border_undefined);
-    /** Initialise the kernel's source, destination and border.
-     *
-     * @note At least one of output_x or output_y must be set.
-     *
-     * @param[in]  compile_context  The compile context to be used.
-     * @param[in]  input            Source tensor. Data types supported: U8.
-     * @param[out] output_x         (Optional) Destination tensor for the X gradient, Data types supported: S32.
-     * @param[out] output_y         (Optional) Destination tensor for the Y gradient, Data types supported: S32.
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output_x, ICLTensor *output_y, bool border_undefined);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-    BorderSize border_size() const override;
-
-private:
-    const ICLTensor *_input;       /**< Input tensor */
-    ICLTensor       *_output_x;    /**< X output of horizontal pass */
-    ICLTensor       *_output_y;    /**< Y output of horizontal pass */
-    bool             _run_sobel_x; /**< Do we need to run Sobel X ? */
-    bool             _run_sobel_y; /**< Do we need to run Sobel Y ? */
-    BorderSize       _border_size; /**< Border size */
-};
-
-/** Interface for the kernel to run the vertical pass of 7x7 Sobel filter on a tensor. */
-class CLSobel7x7VertKernel : public ICLKernel
-{
-public:
-    /** Default constructor: initialize all the pointers to nullptr and parameters to zero. */
-    CLSobel7x7VertKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLSobel7x7VertKernel(const CLSobel7x7VertKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLSobel7x7VertKernel &operator=(const CLSobel7x7VertKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLSobel7x7VertKernel(CLSobel7x7VertKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLSobel7x7VertKernel &operator=(CLSobel7x7VertKernel &&) = default;
-    /** Default destructor */
-    ~CLSobel7x7VertKernel() = default;
-
-    /** Initialise the kernel's source, destination and border.
-     *
-     * @note At least one of output_x or output_y must be set and the corresponding input.
-     *
-     * @param[in]  input_x          (Optional) Input for X (X output of horizontal pass). Data types supported: S32.
-     * @param[in]  input_y          (Optional) Input for Y (Y output of horizontal pass). Data types supported: S32.
-     * @param[out] output_x         (Optional) Destination tensor for the X gradient, Data types supported: S32.
-     * @param[out] output_y         (Optional) Destination tensor for the Y gradient, Data types supported: S32.
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const ICLTensor *input_x, const ICLTensor *input_y, ICLTensor *output_x, ICLTensor *output_y, bool border_undefined);
-    /** Initialise the kernel's source, destination and border.
-     *
-     * @note At least one of output_x or output_y must be set and the corresponding input.
-     *
-     * @param[in]  compile_context  The compile context to be used.
-     * @param[in]  input_x          (Optional) Input for X (X output of horizontal pass). Data types supported: S32.
-     * @param[in]  input_y          (Optional) Input for Y (Y output of horizontal pass). Data types supported: S32.
-     * @param[out] output_x         (Optional) Destination tensor for the X gradient, Data types supported: S32.
-     * @param[out] output_y         (Optional) Destination tensor for the Y gradient, Data types supported: S32.
-     * @param[in]  border_undefined True if the border mode is undefined. False if it's replicate or constant.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input_x, const ICLTensor *input_y, ICLTensor *output_x, ICLTensor *output_y, bool border_undefined);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-    BorderSize border_size() const override;
-
-private:
-    const ICLTensor *_input_x;     /**< X input (X output of the horizontal pass) */
-    const ICLTensor *_input_y;     /**< Y input (Y output of the horizontal pass) */
-    ICLTensor       *_output_x;    /**< X output of sobel */
-    ICLTensor       *_output_y;    /**< Y output of sobel */
-    bool             _run_sobel_x; /**< Do we need to run sobel X? */
-    bool             _run_sobel_y; /**< Do we need to run sobel Y? */
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLSOBEL7X7KERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLSoftmaxLayerKernel.h b/arm_compute/core/CL/kernels/CLSoftmaxLayerKernel.h
deleted file mode 100644
index f8c1019d53..0000000000
--- a/arm_compute/core/CL/kernels/CLSoftmaxLayerKernel.h
+++ /dev/null
@@ -1,158 +0,0 @@
-/*
- * Copyright (c) 2017-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLSOFTMAXLAYERKERNEL_H
-#define ARM_COMPUTE_CLSOFTMAXLAYERKERNEL_H
-
-#include "arm_compute/core/CL/ICLSimple3DKernel.h"
-#include "arm_compute/core/KernelDescriptors.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for max, shifting, exponentiating and summing the logits */
-class CLLogits1DMaxShiftExpSumKernel : public ICLKernel
-{
-public:
-    /** Info for whether a parallel reduction will be run and the vector size of the execution. */
-    using ParallelReductionInfo = std::tuple<bool, unsigned int>;
-
-public:
-    /** Default constructor */
-    CLLogits1DMaxShiftExpSumKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLLogits1DMaxShiftExpSumKernel(const CLLogits1DMaxShiftExpSumKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLLogits1DMaxShiftExpSumKernel &operator=(const CLLogits1DMaxShiftExpSumKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLLogits1DMaxShiftExpSumKernel(CLLogits1DMaxShiftExpSumKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLLogits1DMaxShiftExpSumKernel &operator=(CLLogits1DMaxShiftExpSumKernel &&) = default;
-    /** Set the input and output tensors.
-     *
-     * @param[in]     input  Source tensor. Data types supported: QASYMM8/QASYMM8_SIGNED/F16/F32
-     * @param[in,out] max    Max values tensor. Data types supported: same as @p input
-     * @param[out]    output Destination tensor. Data types supported: same as @p input
-     * @param[out]    sum    Sum of 1D logits tensor. Data types supported: same as @p input
-     * @param[in]     info   Contains information consumed by kernels for softmax described in @ref SoftmaxKernelInfo.
-     */
-    void configure(const ICLTensor *input, ICLTensor *max, ICLTensor *output, ICLTensor *sum, const SoftmaxKernelInfo &info);
-    /** Set the input and output tensors.
-     *
-     * @param[in]     compile_context The compile context to be used.
-     * @param[in]     input           Source tensor. Data types supported: QASYMM8/QASYMM8_SIGNED/F16/F32
-     * @param[in,out] max             Max values tensor. Data types supported: same as @p input
-     * @param[out]    output          Destination tensor. Data types supported: same as @p input
-     * @param[out]    sum             Sum of 1D logits tensor. Data types supported: same as @p input
-     * @param[in]     info            Contains information consumed by kernels for softmax described in @ref SoftmaxKernelInfo.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *max, ICLTensor *output, ICLTensor *sum, const SoftmaxKernelInfo &info);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLLogits1DMaxShiftExpSumKernel
-     *
-     * @param[in] input  Source tensor. Data types supported: QASYMM8/QASYMM8_SIGNED/F16/F32
-     * @param[in] max    Max values tensor. Data types supported: same as @p input
-     * @param[in] output Destination tensor. Data types supported: same as @p input
-     * @param[in] sum    Sum of 1D logits tensor. Data types supported: same as @p input
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *max, const ITensorInfo *output, const ITensorInfo *sum);
-    /** Checks if the given size is eligible for parallel reduction
-     *
-     * @note  Serial reduction is launched for width < (_grid_size * _serial_vector_size).
-     * @note  Parallel reduction is launched for width >= (_grid_size * _serial_vector_size) and vector_size is forced to 4.
-     *
-     * @param[in] size Size to check
-     *
-     * @return A two-element tuple where the first element is a boolean specifying if a parallel reduction will be run,
-     *         while the second element is the vector size of the execution.
-     */
-    static ParallelReductionInfo is_parallel_reduction(size_t size);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;
-    ICLTensor       *_max;
-    ICLTensor       *_output;
-    ICLTensor       *_sum;
-
-private:
-    static const unsigned int _grid_size;
-    static const unsigned int _serial_vector_size;
-    static const unsigned int _parallel_vector_size;
-};
-/** Interface for calculating the final step of the Softmax Layer where each logit value is multiplied by the inverse of the sum of the logits. */
-class CLLogits1DNormKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLLogits1DNormKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLLogits1DNormKernel(const CLLogits1DNormKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLLogits1DNormKernel &operator=(const CLLogits1DNormKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLLogits1DNormKernel(CLLogits1DNormKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLLogits1DNormKernel &operator=(CLLogits1DNormKernel &&) = default;
-    /** Set the input and output tensors.
-     *
-     * @param[in]  input  Source tensor. Data types supported: S32/F16/F32. If this kernel is used for log softmax, only F32/F16 is supported.
-     * @param[in]  sum    Sum tensor. Dimensions should be dim(input)-1. Data types supported: same as @p input
-     * @param[out] output Destination tensor. Data types supported: QASYMM8/QASYMM8_SIGNED for S32 @p input, or same as @p input
-     * @param[in]  info   Contains information consumed by kernels for softmax described in @ref SoftmaxKernelInfo.
-     */
-    void configure(const ICLTensor *input, const ICLTensor *sum, ICLTensor *output, const SoftmaxKernelInfo &info);
-    /** Set the input and output tensors.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Source tensor. Data types supported: S32/F16/F32. If this kernel is used for log softmax, only F32/F16 is supported.
-     * @param[in]  sum             Sum tensor. Dimensions should be dim(input)-1. Data types supported: same as @p input
-     * @param[out] output          Destination tensor. Data types supported: QASYMM8/QASYMM8_SIGNED for S32 @p input, or same as @p input
-     * @param[in]  info            Contains information consumed by kernels for softmax described in @ref SoftmaxKernelInfo.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, const ICLTensor *sum, ICLTensor *output, const SoftmaxKernelInfo &info);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLLogits1DNormKernel
-     *
-     * @param[in] input  Source tensor. Data types supported: S32/F16/F32. If this kernel is used for log softmax, only F32/F16 is supported.
-     * @param[in] sum    Sum tensor. Dimensions should be dim(input)-1. Data types supported: same as @p input
-     * @param[in] output Destination tensor. Data types supported: QASYMM8 for S32 @p input, or same as @p input
-     * @param[in] info   Contains information consumed by kernels for softmax described in @ref SoftmaxKernelInfo.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *sum, const ITensorInfo *output, const SoftmaxKernelInfo &info);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;
-    const ICLTensor *_sum;
-    ICLTensor       *_output;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLSOFTMAXLAYERKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLSpaceToBatchLayerKernel.h b/arm_compute/core/CL/kernels/CLSpaceToBatchLayerKernel.h
deleted file mode 100644
index 93221f7b5a..0000000000
--- a/arm_compute/core/CL/kernels/CLSpaceToBatchLayerKernel.h
+++ /dev/null
@@ -1,121 +0,0 @@
-/*
- * Copyright (c) 2018-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLSPACETOBATCHLAYERKERNEL_H
-#define ARM_COMPUTE_CLSPACETOBATCHLAYERKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-#include "arm_compute/core/Types.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the space to batch kernel */
-class CLSpaceToBatchLayerKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLSpaceToBatchLayerKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLSpaceToBatchLayerKernel(const CLSpaceToBatchLayerKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLSpaceToBatchLayerKernel &operator=(const CLSpaceToBatchLayerKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLSpaceToBatchLayerKernel(CLSpaceToBatchLayerKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLSpaceToBatchLayerKernel &operator=(CLSpaceToBatchLayerKernel &&) = default;
-    /** Default destructor */
-    ~CLSpaceToBatchLayerKernel() = default;
-    /** Initialise the kernel's inputs and output.
-     *
-     * @param[in]  input       Tensor input. Supported tensor rank: 4. Data types supported: All.
-     * @param[in]  block_shape 1-D tensor with shape [M]. Data types supported: S32
-     * @param[in]  paddings    2-D tensor with shape [2, M]. Data types supported: S32
-     * @param[out] output      Tensor output. Data types supported: same as @p input
-     */
-    void configure(const ICLTensor *input, const ICLTensor *block_shape, const ICLTensor *paddings, ICLTensor *output);
-    /** Initialise the kernel's inputs and output.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Tensor input. Supported tensor rank: 4. Data types supported: All.
-     * @param[in]  block_shape     1-D tensor with shape [M]. Data types supported: S32
-     * @param[in]  paddings        2-D tensor with shape [2, M]. Data types supported: S32
-     * @param[out] output          Tensor output. Data types supported: same as @p input
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, const ICLTensor *block_shape, const ICLTensor *paddings, ICLTensor *output);
-    /** Initialise the kernel's input and output. (Static block shape and paddings)
-     *
-     * @param[in]  input         Tensor input. Supported tensor rank: 4. Data types supported: All.
-     * @param[in]  block_shape_x Block shape x value.
-     * @param[in]  block_shape_y Block shape y value.
-     * @param[in]  padding_left  The left padding of the output tensor.
-     * @param[in]  padding_right The right padding of the output tensor.
-     * @param[out] output        Tensor output. Data types supported: same as @p input
-     */
-    void configure(const ICLTensor *input, const int block_shape_x, const int block_shape_y, const Size2D &padding_left, const Size2D &padding_right, ICLTensor *output);
-    /** Initialise the kernel's input and output. (Static block shape and paddings)
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Tensor input. Supported tensor rank: 4. Data types supported: All.
-     * @param[in]  block_shape_x   Block shape x value.
-     * @param[in]  block_shape_y   Block shape y value.
-     * @param[in]  padding_left    The left padding of the output tensor.
-     * @param[in]  padding_right   The right padding of the output tensor.
-     * @param[out] output          Tensor output. Data types supported: same as @p input
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, const int block_shape_x, const int block_shape_y, const Size2D &padding_left, const Size2D &padding_right, ICLTensor *output);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLSpaceToBatchLayerKernel
-     *
-     * @param[in] input       Tensor input. Supported tensor rank: 4. Data types supported: All.
-     * @param[in] block_shape 1-D tensor with shape [M]. Data types supported: S32
-     * @param[in] paddings    2-D tensor with shape [2, M]. Data types supported: S32
-     * @param[in] output      Tensor output. Data types supported: same as @p input
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *block_shape, const ITensorInfo *paddings, const ITensorInfo *output);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLSpaceToBatchLayerKernel (Static block shape and paddings)
-     *
-     * @param[in] input         Tensor input. Supported tensor rank: 4. Data types supported: All.
-     * @param[in] block_shape_x Block shape x value.
-     * @param[in] block_shape_y Block shape y value.
-     * @param[in] padding_left  The left padding of the output tensor.
-     * @param[in] padding_right The right padding of the output tensor.
-     * @param[in] output        Tensor output. Data types supported: same as @p input
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const int block_shape_x, const int block_shape_y, const Size2D &padding_left, const Size2D &padding_right, const ITensorInfo *output);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;       /**< Source tensor */
-    const ICLTensor *_block_shape; /**< Block shape tensor */
-    const ICLTensor *_paddings;    /**< Paddings tensor */
-    ICLTensor       *_output;      /**< Destination tensor */
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLSPACETOBATCHLAYERKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLSpaceToDepthLayerKernel.h b/arm_compute/core/CL/kernels/CLSpaceToDepthLayerKernel.h
deleted file mode 100644
index af0aa12598..0000000000
--- a/arm_compute/core/CL/kernels/CLSpaceToDepthLayerKernel.h
+++ /dev/null
@@ -1,84 +0,0 @@
-/*
- * Copyright (c) 2019-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLSPACETODEPTHLAYERKERNEL_H
-#define ARM_COMPUTE_CLSPACETODEPTHLAYERKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-#include "arm_compute/core/Types.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the space to depth kernel */
-class CLSpaceToDepthLayerKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLSpaceToDepthLayerKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLSpaceToDepthLayerKernel(const CLSpaceToDepthLayerKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLSpaceToDepthLayerKernel &operator=(const CLSpaceToDepthLayerKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLSpaceToDepthLayerKernel(CLSpaceToDepthLayerKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLSpaceToDepthLayerKernel &operator=(CLSpaceToDepthLayerKernel &&) = default;
-    /** Default destructor */
-    ~CLSpaceToDepthLayerKernel() = default;
-    /** Initialise the kernel's inputs and output.
-     *
-     * @param[in]  input       Tensor input. Supported tensor rank: 4. Data types supported: All.
-     * @param[out] output      Tensor output. Data types supported: same as @p input
-     * @param[in]  block_shape Block shape value.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, int32_t block_shape);
-    /** Initialise the kernel's inputs and output.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Tensor input. Supported tensor rank: 4. Data types supported: All.
-     * @param[out] output          Tensor output. Data types supported: same as @p input
-     * @param[in]  block_shape     Block shape value.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, int32_t block_shape);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLSpaceToDepthLayerKernel.
-     *
-     * @param[in] input       Tensor input info. Supported tensor rank: 4. Data types supported: All.
-     * @param[in] output      Tensor output info. Data types supported: same as @p input
-     * @param[in] block_shape Block shape value.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output, int32_t block_shape);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;       /**< Source tensor */
-    ICLTensor       *_output;      /**< Destination tensor */
-    int32_t          _block_shape; /**< Block shape */
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLSPACETODEPTHLAYERKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLStackLayerKernel.h b/arm_compute/core/CL/kernels/CLStackLayerKernel.h
deleted file mode 100644
index cfefcd97dd..0000000000
--- a/arm_compute/core/CL/kernels/CLStackLayerKernel.h
+++ /dev/null
@@ -1,101 +0,0 @@
-/*
- * Copyright (c) 2018-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-
-#ifndef ARM_COMPUTE_CLSTACKLAYERKERNEL_H
-#define ARM_COMPUTE_CLSTACKLAYERKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-#include "arm_compute/core/Types.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** OpenCL kernel to stacks a rank-R tensor into one with rank-(R+1) along the axis dimension.*/
-class CLStackLayerKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLStackLayerKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLStackLayerKernel(const CLStackLayerKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLStackLayerKernel &operator=(const CLStackLayerKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLStackLayerKernel(CLStackLayerKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLStackLayerKernel &operator=(CLStackLayerKernel &&) = default;
-    /** Default destructor */
-    ~CLStackLayerKernel() = default;
-    /** Initialise the kernel's inputs and output
-     *
-     * @note Supported input tensor rank: up to 4
-     *
-     * @param[in]  input       Input tensor. Data types supported: All.
-     * @param[in]  axis        The dimension to stack the tensors along. It must be smaller than the number of input dimensions.
-     * @param[in]  idx_input   Index of the input tensor in the list of tensors to stack.
-     *                         All tensors in the list must have the same shape
-     * @param[in]  num_tensors Number of tensors to stack
-     * @param[out] output      Output tensor. Data types supported: Same as @p input.
-     *
-     */
-    void configure(const ICLTensor *input, unsigned int axis, unsigned int idx_input, unsigned int num_tensors, ICLTensor *output);
-    /** Initialise the kernel's inputs and output
-     *
-     * @note Supported input tensor rank: up to 4
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Input tensor. Data types supported: All.
-     * @param[in]  axis            The dimension to stack the tensors along. It must be smaller than the number of input dimensions.
-     * @param[in]  idx_input       Index of the input tensor in the list of tensors to stack.
-     *                             All tensors in the list must have the same shape
-     * @param[in]  num_tensors     Number of tensors to stack
-     * @param[out] output          Output tensor. Data types supported: Same as @p input.
-     *
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, unsigned int axis, unsigned int idx_input, unsigned int num_tensors, ICLTensor *output);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLStackLayerKernel
-     *
-     * @note Supported input tensor rank: up to 4
-     *
-     * @param[in] input       Input tensor info. Data types supported: All.
-     * @param[in] axis        The dimension to stack the tensors along. It must be smaller than the number of input dimensions.
-     * @param[in] idx_input   Index of the input tensor in the list of tensors to stack
-     *                        All tensors in the list must have the same shape
-     * @param[in] num_tensors Number of tensors to stack
-     * @param[in] output      Output tensor info. Data types supported: Same as @p input.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, unsigned int axis, unsigned int idx_input, unsigned int num_tensors, const ITensorInfo *output);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;
-    ICLTensor       *_output;
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLSTACKLAYERKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLStridedSliceKernel.h b/arm_compute/core/CL/kernels/CLStridedSliceKernel.h
deleted file mode 100644
index 74311b71fa..0000000000
--- a/arm_compute/core/CL/kernels/CLStridedSliceKernel.h
+++ /dev/null
@@ -1,79 +0,0 @@
-/*
- * Copyright (c) 2018-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CL_STRIDED_SLICE_KERNEL_H
-#define ARM_COMPUTE_CL_STRIDED_SLICE_KERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-#include "arm_compute/core/Types.h"
-
-#include <cstdint>
-
-namespace arm_compute
-{
-/** Interface for the kernel to perform tensor strided slicing */
-class CLStridedSliceKernel : public ICLKernel
-{
-public:
-    /** Configure kernel
-     *
-     * @note Supported tensor rank: up to 4
-     *
-     * @param[in]  compile_context  The compile context to be used.
-     * @param[in]  input            Source tensor info. Data type supported: All.
-     * @param[out] output           Destination tensor info. Data type supported: Same as @p input
-     * @param[in]  starts           The starts of the dimensions of the input tensor to be sliced. The length must be of rank(input).
-     * @param[in]  ends             The ends of the dimensions of the input tensor to be sliced. The length must be of rank(input).
-     * @param[in]  strides          The strides of the dimensions of the input tensor to be sliced. The length must be of rank(input).
-     * @param[in]  begin_mask       If the ith bit of begin_mask is set, starts[i] is ignored and the fullest possible range in that dimension is used instead.
-     * @param[in]  end_mask         If the ith bit of end_mask is set, ends[i] is ignored and the fullest possible range in that dimension is used instead.
-     * @param[in]  shrink_axis_mask If the ith bit of shrink_axis_mask is set, it implies that the ith specification shrinks the dimensionality by 1.
-     *                              A slice of size 1 starting from starts[i] in the dimension must be preserved.
-     */
-    void configure(const CLCompileContext &compile_context, const ITensorInfo *input, ITensorInfo *output,
-                   const Coordinates &starts, const Coordinates &ends, const BiStrides &strides,
-                   int32_t begin_mask, int32_t end_mask, int32_t shrink_axis_mask);
-
-    /** Static function to check if given info will lead to a valid configuration of @ref CLStridedSliceKernel
-     *
-     * @note Supported tensor rank: up to 4
-     *
-     * @param[in] input            Source tensor. Data type supported: All.
-     * @param[in] output           Destination tensor. Data type supported: Same as @p input
-     * @param[in] starts           The starts of the dimensions of the input tensor to be sliced. The length must be of rank(input).
-     * @param[in] ends             The ends of the dimensions of the input tensor to be sliced. The length must be of rank(input).
-     * @param[in] strides          The strides of the dimensions of the input tensor to be sliced. The length must be of rank(input).
-     * @param[in] begin_mask       If the ith bit of begin_mask is set, starts[i] is ignored and the fullest possible range in that dimension is used instead.
-     * @param[in] end_mask         If the ith bit of end_mask is set, ends[i] is ignored and the fullest possible range in that dimension is used instead.
-     * @param[in] shrink_axis_mask If the ith bit of shrink_axis_mask is set, it implies that the ith specification shrinks the dimensionality by 1.
-     *                             A slice of size 1 starting from starts[i] in the dimension must be preserved.
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output,
-                           const Coordinates &starts, const Coordinates &ends, const BiStrides &strides,
-                           int32_t begin_mask, int32_t end_mask, int32_t shrink_axis_mask);
-
-    // Inherited methods overridden:
-    void run_op(ITensorPack &tensors, const Window &window, cl::CommandQueue &queue) override;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CL_STRIDED_SLICE_KERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLTableLookupKernel.h b/arm_compute/core/CL/kernels/CLTableLookupKernel.h
deleted file mode 100644
index 9f1d28c47a..0000000000
--- a/arm_compute/core/CL/kernels/CLTableLookupKernel.h
+++ /dev/null
@@ -1,55 +0,0 @@
-/*
- * Copyright (c) 2016-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLTABLELOOKUPKERNEL_H
-#define ARM_COMPUTE_CLTABLELOOKUPKERNEL_H
-
-#include "arm_compute/core/CL/ICLSimple2DKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-class ICLLut;
-
-/** Interface for the kernel to perform table lookup calculations. */
-class CLTableLookupKernel : public ICLSimple2DKernel
-{
-public:
-    /** Initialise the kernel's input, lut and output.
-     *
-     * @param[in]  input  An input tensor. Data types supported: U8, S16.
-     * @param[in]  lut    The input LUT. Data types supported: U8, S16.
-     * @param[out] output The output tensor. Data types supported: U8, S16.
-     */
-    void configure(const ICLTensor *input, const ICLLut *lut, ICLTensor *output);
-    /** Initialise the kernel's input, lut and output.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           An input tensor. Data types supported: U8, S16.
-     * @param[in]  lut             The input LUT. Data types supported: U8, S16.
-     * @param[out] output          The output tensor. Data types supported: U8, S16.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, const ICLLut *lut, ICLTensor *output);
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLTABLELOOKUPKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLThresholdKernel.h b/arm_compute/core/CL/kernels/CLThresholdKernel.h
deleted file mode 100644
index 7e01fd6aaa..0000000000
--- a/arm_compute/core/CL/kernels/CLThresholdKernel.h
+++ /dev/null
@@ -1,57 +0,0 @@
-/*
- * Copyright (c) 2016-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLTHRESHOLDKERNEL_H
-#define ARM_COMPUTE_CLTHRESHOLDKERNEL_H
-
-#include "arm_compute/core/CL/ICLSimple2DKernel.h"
-#include "arm_compute/core/KernelDescriptors.h"
-#include "arm_compute/core/Types.h"
-
-namespace arm_compute
-{
-// Forward declarations
-class ICLTensor;
-
-/** Interface for the thresholding kernel. */
-class CLThresholdKernel : public ICLSimple2DKernel
-{
-public:
-    /**Initialise the kernel's input, output and threshold parameters.
-     *
-     * @param[in]  input  An input tensor. Data types supported: U8
-     * @param[out] output The output tensor. Data types supported: U8.
-     * @param[in]  info   Threshold descriptor
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, const ThresholdKernelInfo &info);
-    /**Initialise the kernel's input, output and threshold parameters.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           An input tensor. Data types supported: U8
-     * @param[out] output          The output tensor. Data types supported: U8.
-     * @param[in]  info            Threshold descriptor
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, const ThresholdKernelInfo &info);
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_NETHRESHOLDKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLTileKernel.h b/arm_compute/core/CL/kernels/CLTileKernel.h
deleted file mode 100644
index 56e1df8de3..0000000000
--- a/arm_compute/core/CL/kernels/CLTileKernel.h
+++ /dev/null
@@ -1,88 +0,0 @@
-/*
- * Copyright (c) 2018-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLTILEKERNEL_H
-#define ARM_COMPUTE_CLTILEKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** OpenCL kernel to perform a Tile operation */
-class CLTileKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLTileKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLTileKernel(const CLTileKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLTileKernel &operator=(const CLTileKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLTileKernel(CLTileKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLTileKernel &operator=(CLTileKernel &&) = default;
-    /** Default destructor */
-    ~CLTileKernel() = default;
-    /** Set the source, destination of the kernel
-     *
-     * @param[in]  input     Source tensor. Data type supported: All.
-     * @param[in]  multiples Contains the number of times the input tensor should be replicated on the given dimension.
-     *                       Cannot have more than 4 elements (tiling in dimensions greater than 4 is not supported).
-     * @param[out] output    Destination tensor. Same as @p input
-     *
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, const Multiples &multiples);
-    /** Set the source, destination of the kernel
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Source tensor. Data type supported: All.
-     * @param[in]  multiples       Contains the number of times the input tensor should be replicated on the given dimension.
-     *                             Cannot have more than 4 elements (tiling in dimensions greater than 4 is not supported).
-     * @param[out] output          Destination tensor. Same as @p input
-     *
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, const Multiples &multiples);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLTileKernel
-     *
-     * @param[in] input     Source tensor info. Data type supported: All.
-     * @param[in] multiples Contains the number of times the input tensor should be replicated on the given dimension.
-     *                      Cannot have more than 4 elements (tiling in dimensions greater than 4 is not supported).
-     * @param[in] output    Destination tensor info. Same as @p input
-     *
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output, const Multiples &multiples);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;
-    ICLTensor       *_output;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLTILEKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLTransposeKernel.h b/arm_compute/core/CL/kernels/CLTransposeKernel.h
deleted file mode 100644
index 4a9887f2cf..0000000000
--- a/arm_compute/core/CL/kernels/CLTransposeKernel.h
+++ /dev/null
@@ -1,64 +0,0 @@
-/*
- * Copyright (c) 2017-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLTRANSPOSEKERNEL_H
-#define ARM_COMPUTE_CLTRANSPOSEKERNEL_H
-
-#include "arm_compute/core/CL/ICLSimple2DKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** OpenCL kernel which transposes the elements of a matrix.
- *
- * [width, height, batch] -> [height, width, batch]
- *
- */
-class CLTransposeKernel : public ICLSimple2DKernel
-{
-public:
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in]  input  Input tensor. Data types supported: All.
-     * @param[out] output Output tensor. Data type supported: Same as @p input
-     */
-    void configure(const ICLTensor *input, ICLTensor *output);
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Input tensor. Data types supported: All.
-     * @param[out] output          Output tensor. Data type supported: Same as @p input
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLTransposeKernel
-     *
-     * @param[in] input  Input tensor. Data types supported: All.
-     * @param[in] output Output tensor. Data type supported: Same as @p input
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output);
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLTRANSPOSEKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLUpsampleLayerKernel.h b/arm_compute/core/CL/kernels/CLUpsampleLayerKernel.h
deleted file mode 100644
index b523b97233..0000000000
--- a/arm_compute/core/CL/kernels/CLUpsampleLayerKernel.h
+++ /dev/null
@@ -1,89 +0,0 @@
-/*
- * Copyright (c) 2018-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLUPSAMPLELAYERKERNEL_H
-#define ARM_COMPUTE_CLUPSAMPLELAYERKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the UpsampleLayer kernel on OpenCL. */
-class CLUpsampleLayerKernel : public ICLKernel
-{
-public:
-    /** Constructor */
-    CLUpsampleLayerKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLUpsampleLayerKernel(const CLUpsampleLayerKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLUpsampleLayerKernel &operator=(const CLUpsampleLayerKernel &) = delete;
-    /** Default Move Constructor. */
-    CLUpsampleLayerKernel(CLUpsampleLayerKernel &&) = default;
-    /** Default move assignment operator */
-    CLUpsampleLayerKernel &operator=(CLUpsampleLayerKernel &&) = default;
-    /** Default destructor */
-    ~CLUpsampleLayerKernel() = default;
-
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in]  input             Source tensor. Data types supported: All.
-     * @param[out] output            Destination tensor. Data types supported: same as @p input.
-     * @param[in]  info              Contains stride information described in @ref Size2D.
-     * @param[in]  upsampling_policy Defines the policy to fill the intermediate pixels.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, const Size2D &info, const InterpolationPolicy upsampling_policy);
-    /** Initialise the kernel's input and output.
-     *
-     * @param[in]  compile_context   The compile context to be used.
-     * @param[in]  input             Source tensor. Data types supported: All.
-     * @param[out] output            Destination tensor. Data types supported: same as @p input.
-     * @param[in]  info              Contains stride information described in @ref Size2D.
-     * @param[in]  upsampling_policy Defines the policy to fill the intermediate pixels.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, const Size2D &info, const InterpolationPolicy upsampling_policy);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLUpsampleLayerKernel
-     *
-     * @param[in] input             Source tensor info. Data types supported: All.
-     * @param[in] output            Destination tensor info. Data types supported: same as @p input.
-     * @param[in] info              Contains  stride information described in @ref Size2D.
-     * @param[in] upsampling_policy Defines the policy to fill the intermediate pixels.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output, const Size2D &info, const InterpolationPolicy upsampling_policy);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;
-    ICLTensor       *_output;
-    Size2D           _info;
-    DataLayout       _data_layout;
-    unsigned int     _num_elems_processed_per_iteration_input_x;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLUPSAMPLELAYERKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLWarpAffineKernel.h b/arm_compute/core/CL/kernels/CLWarpAffineKernel.h
deleted file mode 100644
index 440febab96..0000000000
--- a/arm_compute/core/CL/kernels/CLWarpAffineKernel.h
+++ /dev/null
@@ -1,62 +0,0 @@
-/*
- * Copyright (c) 2016-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLWARPAFFINEKERNEL_H
-#define ARM_COMPUTE_CLWARPAFFINEKERNEL_H
-
-#include "arm_compute/core/CL/ICLSimple2DKernel.h"
-#include "arm_compute/core/Types.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the warp affine kernel.*/
-class CLWarpAffineKernel : public ICLSimple2DKernel
-{
-public:
-    /** Initialize the function's source, destination, interpolation policy and border_mode.
-     *
-     * @param[in]  input  Source tensor. Data types supported: U8.
-     * @param[out] output Destination tensor, Data types supported: U8.
-     * @param[in]  matrix The perspective matrix. Must be 2x3 of type float
-     *                    The matrix argument requires 9 values, the last 3 values are ignored.
-     * @param[in]  policy The interpolation type.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, const std::array<float, 9> &matrix, InterpolationPolicy policy);
-    /** Initialize the function's source, destination, interpolation policy and border_mode.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Source tensor. Data types supported: U8.
-     * @param[out] output          Destination tensor, Data types supported: U8.
-     * @param[in]  matrix          The perspective matrix. Must be 2x3 of type float
-     *                             The matrix argument requires 9 values, the last 3 values are ignored.
-     * @param[in]  policy          The interpolation type.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, const std::array<float, 9> &matrix, InterpolationPolicy policy);
-
-    // Inherited methods overridden:
-    BorderSize border_size() const override;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLWARPAFFINEKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLWarpPerspectiveKernel.h b/arm_compute/core/CL/kernels/CLWarpPerspectiveKernel.h
deleted file mode 100644
index 6614989059..0000000000
--- a/arm_compute/core/CL/kernels/CLWarpPerspectiveKernel.h
+++ /dev/null
@@ -1,59 +0,0 @@
-/*
- * Copyright (c) 2016-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLWARPERSPECTIVEKERNEL_H
-#define ARM_COMPUTE_CLWARPERSPECTIVEKERNEL_H
-
-#include "arm_compute/core/CL/ICLSimple2DKernel.h"
-#include "arm_compute/core/Types.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-/** Interface for the warp perspective kernel.*/
-class CLWarpPerspectiveKernel : public ICLSimple2DKernel
-{
-public:
-    /** Initialize the function's source, destination, interpolation policy and border_mode.
-     *
-     * @param[in]  input  Source tensor. Data types supported: U8.
-     * @param[out] output Destination tensor, Data types supported: U8.
-     * @param[in]  matrix The perspective matrix. Must be 3x3 of type float.
-     * @param[in]  policy The interpolation type.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, const std::array<float, 9> &matrix, InterpolationPolicy policy);
-    /** Initialize the function's source, destination, interpolation policy and border_mode.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Source tensor. Data types supported: U8.
-     * @param[out] output          Destination tensor, Data types supported: U8.
-     * @param[in]  matrix          The perspective matrix. Must be 3x3 of type float.
-     * @param[in]  policy          The interpolation type.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, const std::array<float, 9> &matrix, InterpolationPolicy policy);
-
-    // Inherited methods overridden:
-    BorderSize border_size() const override;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLWARPERSPECTIVEKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLWeightsReshapeKernel.h b/arm_compute/core/CL/kernels/CLWeightsReshapeKernel.h
deleted file mode 100644
index c74255bac0..0000000000
--- a/arm_compute/core/CL/kernels/CLWeightsReshapeKernel.h
+++ /dev/null
@@ -1,121 +0,0 @@
-/*
- * Copyright (c) 2017-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLWEIGHTSRESHAPEKERNEL_H
-#define ARM_COMPUTE_CLWEIGHTSRESHAPEKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-/** OpenCL kernel to perform reshaping on the weights used by convolution and locally connected layer
- *
- * Rearranges each 3-dimensional kernel to a single row leading to a matrix with linearized kernels.
- * In combination with the @ref CLIm2ColKernel can transform a convolution to a matrix multiplication.
- *
- * For example assuming a 3D weight kernel of 3x3 dimensions and depth of 2 we have:
- * @f[
- * \left( \begin{array}{ccc}
- * a000 & a001 & a002 \\
- * a010 & a011 & a012 \\
- * a020 & a021 & a022 \\
- * \end{array} \right)
- * \left( \begin{array}{ccc}
- * a100 & a101 & a102 \\
- * a110 & a111 & a112 \\
- * a120 & a121 & a122 \\
- * \end{array} \right)
- * \rightarrow
- * \left( \begin{array}{ccccccccc}
- * a000 & a001 & a002 & a010 & a011 & a012 & a020 & a021 & a022 & a100 & a101 & a102 & a110 & a111 & a112 & a120 & a121 & a122 \\
- * \end{array} \right)
- * @f]
- */
-class CLWeightsReshapeKernel : public ICLKernel
-{
-public:
-    /** Constructor.*/
-    CLWeightsReshapeKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLWeightsReshapeKernel(const CLWeightsReshapeKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLWeightsReshapeKernel &operator=(const CLWeightsReshapeKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLWeightsReshapeKernel(CLWeightsReshapeKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLWeightsReshapeKernel &operator=(CLWeightsReshapeKernel &&) = default;
-    /** Default destructor */
-    ~CLWeightsReshapeKernel() = default;
-    /** Set the input and output of the kernel.
-     *
-     * @param[in]  input      The input tensor to convert. Weights are 4D tensor with dimensions [kernel_x, kernel_y, IFM, OFM] if shared,
-     *                        and 5D tensor with dimensions [kernel_x, kernel_y, IFM, OFM,  num_patches] if unshared. Data types supported: All
-     * @param[in]  biases     The shared biases tensor to append.  Bias is 1D tensor with dimensions [OFM] if shared and 2D tensor with
-     *                        dimensions [OFM, num_patches] if unshared. Data types supported: F16/F32, for quantized types this must be nullptr.
-     *                        @warning Appending biases to weights reshaped matrix is not supported for quantized asymmetric types.
-     * @param[out] output     The output tensor. Should be a 2D Tensor if there are no groups and the weights are not shared; a 3D Tensor otherwise.
-     *                        Data types supported: Same as @p input
-     * @param[in]  num_groups (Optional) Number of groups when performing a grouped convolution. num_groups != 1 is only supported for NCHW data layout
-     *                        Number of groups greater than one are only supported for NCHW data layout, and the number of weights must be a multiple of it.
-     */
-    void configure(const ICLTensor *input, const ICLTensor *biases, ICLTensor *output, unsigned int num_groups = 1);
-    /** Set the input and output of the kernel.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           The input tensor to convert. Weights are 4D tensor with dimensions [kernel_x, kernel_y, IFM, OFM] if shared,
-     *                             and 5D tensor with dimensions [kernel_x, kernel_y, IFM, OFM,  num_patches] if unshared. Data types supported: All
-     * @param[in]  biases          The shared biases tensor to append.  Bias is 1D tensor with dimensions [OFM] if shared and 2D tensor with
-     *                             dimensions [OFM, num_patches] if unshared. Data types supported: F16/F32, for quantized types this must be nullptr.
-     *                             @warning Appending biases to weights reshaped matrix is not supported for quantized asymmetric types.
-     * @param[out] output          The output tensor. Should be a 2D Tensor if there are no groups and the weights are not shared; a 3D Tensor otherwise.
-     *                             Data types supported: Same as @p input
-     * @param[in]  num_groups      (Optional) Number of groups when performing a grouped convolution. num_groups != 1 is only supported for NCHW data layout
-     *                             Number of groups greater than one are only supported for NCHW data layout, and the number of weights must be a multiple of it.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, const ICLTensor *biases, ICLTensor *output, unsigned int num_groups = 1);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLWeightsReshapeKernel
-     *
-     * @param[in] input      The input tensor to convert. Weights are 4D tensor with dimensions [kernel_x, kernel_y, IFM, OFM] if shared,
-     *                       and 5D tensor with dimensions [kernel_x, kernel_y, IFM, OFM,  num_patches] if unshared. Data types supported: All
-     * @param[in] biases     The shared biases tensor to append.  Bias is 1D tensor with dimensions [OFM] if shared and 2D tensor with
-     *                       dimensions [OFM, num_patches] if unshared. Data types supported: F16/F32, for quantized types this must be nullptr.
-     *                       @warning Appending biases to weights reshaped matrix is not supported for quantized asymmetric types.
-     * @param[in] output     The output tensor. Should be a 2D Tensor if there are no groups and the weights are not shared; a 3D Tensor otherwise.
-     *                       Data types supported: Same as @p input
-     * @param[in] num_groups (Optional) Number of groups when performing a grouped convolution. num_groups != 1 is only supported for NCHW data layout
-     *                       Number of groups greater than one are only supported for NCHW data layout, and the number of weights must be a multiple of it.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *biases, const ITensorInfo *output, unsigned int num_groups = 1);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;
-    const ICLTensor *_biases;
-    ICLTensor       *_output;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLWEIGHTSRESHAPEKERNEL_H */
\ No newline at end of file
diff --git a/arm_compute/core/CL/kernels/CLWidthConcatenate2TensorsKernel.h b/arm_compute/core/CL/kernels/CLWidthConcatenate2TensorsKernel.h
deleted file mode 100644
index a379b5f0b8..0000000000
--- a/arm_compute/core/CL/kernels/CLWidthConcatenate2TensorsKernel.h
+++ /dev/null
@@ -1,73 +0,0 @@
-/*
- * Copyright (c) 2018-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-
-#ifndef ARM_COMPUTE_CLWIDTHCONCATENATE_2TENSORS_KERNEL_H
-#define ARM_COMPUTE_CLWIDTHCONCATENATE_2TENSORS_KERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-#include "arm_compute/core/Types.h"
-
-namespace arm_compute
-{
-/** Interface for the width concatenate kernel of 2 tensors.
- *  The input1 and input2 tensors will be concatenated into the output tensor.
- */
-class CLWidthConcatenate2TensorsKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLWidthConcatenate2TensorsKernel() = default;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLWidthConcatenate2TensorsKernel(const CLWidthConcatenate2TensorsKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLWidthConcatenate2TensorsKernel &operator=(const CLWidthConcatenate2TensorsKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLWidthConcatenate2TensorsKernel(CLWidthConcatenate2TensorsKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLWidthConcatenate2TensorsKernel &operator=(CLWidthConcatenate2TensorsKernel &&) = default;
-    /** Default destructor */
-    ~CLWidthConcatenate2TensorsKernel() = default;
-    /** Initialise the kernel's input1s and output
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input1          First input tensor. Data types supported: All.
-     * @param[in]  input2          Second input tensor. Data types supported: same as @p input1
-     * @param[out] output          Output tensor. Data types supported: Same as @p input1.
-     */
-    void configure(const CLCompileContext &compile_context, ITensorInfo *input1, ITensorInfo *input2, ITensorInfo *output);
-    /**  Static function to check if given info will lead to a valid configuration of @ref CLWidthConcatenate2TensorsKernel
-     *
-     * @param[in] input1 First tensor info. Data types supported: All.
-     * @param[in] input2 Second tensor info. Data types supported: same as @p input1
-     * @param[in] output Output tensor info. Data types supported: Same as @p input1.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input1, const ITensorInfo *input2, const ITensorInfo *output);
-
-    // Inherited methods overridden:
-    void run_op(ITensorPack &tensors, const Window &window, cl::CommandQueue &queue) override;
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLWIDTHCONCATENATE_2TENSORS_KERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLWidthConcatenate4TensorsKernel.h b/arm_compute/core/CL/kernels/CLWidthConcatenate4TensorsKernel.h
deleted file mode 100644
index 6b0e8ee21d..0000000000
--- a/arm_compute/core/CL/kernels/CLWidthConcatenate4TensorsKernel.h
+++ /dev/null
@@ -1,77 +0,0 @@
-/*
- * Copyright (c) 2018-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-
-#ifndef ARM_COMPUTE_CLWIDTHCONCATENATE_4TENSORS_KERNEL_H
-#define ARM_COMPUTE_CLWIDTHCONCATENATE_4TENSORS_KERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-#include "arm_compute/core/Types.h"
-
-namespace arm_compute
-{
-/** Interface for the width concatenate kernel of 4 tensors.
- *  All input tensors will be concatenated into the output tensor.
- */
-class CLWidthConcatenate4TensorsKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLWidthConcatenate4TensorsKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLWidthConcatenate4TensorsKernel(const CLWidthConcatenate4TensorsKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLWidthConcatenate4TensorsKernel &operator=(const CLWidthConcatenate4TensorsKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLWidthConcatenate4TensorsKernel(CLWidthConcatenate4TensorsKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLWidthConcatenate4TensorsKernel &operator=(CLWidthConcatenate4TensorsKernel &&) = default;
-    /** Default destructor */
-    ~CLWidthConcatenate4TensorsKernel() = default;
-    /** Initialise the kernel's input1s and output
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input1          First input tensor. Data types supported: All.
-     * @param[in]  input2          Second input tensor. Data types supported: same as @p input1
-     * @param[in]  input3          Third input tensor. Data types supported: same as @p input1
-     * @param[in]  input4          Fourth input tensor. Data types supported: same as @p input1
-     * @param[out] output          Output tensor. Data types supported: Same as @p input1.
-     */
-    void configure(const CLCompileContext &compile_context, ITensorInfo *input1, ITensorInfo *input2, ITensorInfo *input3, ITensorInfo *input4, ITensorInfo *output);
-    /**  Static function to check if given info will lead to a valid configuration of @ref CLWidthConcatenate4TensorsKernel
-     *
-     * @param[in] input1 First tensor info. Data types supported: All.
-     * @param[in] input2 Second tensor info. Data types supported: same as @p input1
-     * @param[in] input3 Third tensor info. Data types supported: same as @p input1
-     * @param[in] input4 Fourth tensor info. Data types supported: same as @p input1
-     * @param[in] output Output tensor info. Data types supported: Same as @p input1.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input1, const ITensorInfo *input2, const ITensorInfo *input3, const ITensorInfo *input4, const ITensorInfo *output);
-
-    // Inherited methods overridden:
-    void run_op(ITensorPack &tensors, const Window &window, cl::CommandQueue &queue) override;
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLWIDTHCONCATENATE_4TENSORS_KERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLWidthConcatenateLayerKernel.h b/arm_compute/core/CL/kernels/CLWidthConcatenateLayerKernel.h
deleted file mode 100644
index 16cf167b25..0000000000
--- a/arm_compute/core/CL/kernels/CLWidthConcatenateLayerKernel.h
+++ /dev/null
@@ -1,74 +0,0 @@
-/*
- * Copyright (c) 2018-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-
-#ifndef ARM_COMPUTE_CLWIDTHCONCATENATELAYERKERNEL_H
-#define ARM_COMPUTE_CLWIDTHCONCATENATELAYERKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-#include "arm_compute/core/Types.h"
-
-namespace arm_compute
-{
-/** Interface for the width concatenate kernel.
- *  The input tensor will be concatenated into the output tensor.
- */
-class CLWidthConcatenateLayerKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLWidthConcatenateLayerKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLWidthConcatenateLayerKernel(const CLWidthConcatenateLayerKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLWidthConcatenateLayerKernel &operator=(const CLWidthConcatenateLayerKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLWidthConcatenateLayerKernel(CLWidthConcatenateLayerKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLWidthConcatenateLayerKernel &operator=(CLWidthConcatenateLayerKernel &&) = default;
-    /** Default destructor */
-    ~CLWidthConcatenateLayerKernel() = default;
-    /** Initialise the kernel's inputs and output
-     *
-     * @param[in]     compile_context The compile context to be used.
-     * @param[in]     input           Input tensor. Data types supported: All.
-     * @param[in]     width_offset    The offset on the X axis.
-     * @param[in,out] output          Output tensor. Data types supported: Same as @p input.
-     *
-     */
-    void configure(const CLCompileContext &compile_context, ITensorInfo *input, unsigned int width_offset, ITensorInfo *output);
-    /**  Static function to check if given info will lead to a valid configuration of @ref CLWidthConcatenateLayerKernel
-     *
-     * @param[in] input        Input tensor info. Data types supported: All.
-     * @param[in] width_offset The offset on the X axis.
-     * @param[in] output       Output tensor info. Data types supported: Same as @p input.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, unsigned int width_offset, const ITensorInfo *output);
-
-    // Inherited methods overridden:
-    void run_op(ITensorPack &tensors, const Window &window, cl::CommandQueue &queue) override;
-};
-} // namespace arm_compute
-#endif /* ARM_COMPUTE_CLWIDTHCONCATENATELAYERKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLWinogradFilterTransformKernel.h b/arm_compute/core/CL/kernels/CLWinogradFilterTransformKernel.h
deleted file mode 100644
index b689be820f..0000000000
--- a/arm_compute/core/CL/kernels/CLWinogradFilterTransformKernel.h
+++ /dev/null
@@ -1,115 +0,0 @@
-/*
- * Copyright (c) 2018-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLWINOGRADFILTERTRANSFORMKERNEL_H
-#define ARM_COMPUTE_CLWINOGRADFILTERTRANSFORMKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the Winograd filter transform kernel. */
-class CLWinogradFilterTransformKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLWinogradFilterTransformKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLWinogradFilterTransformKernel(const CLWinogradFilterTransformKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLWinogradFilterTransformKernel &operator=(const CLWinogradFilterTransformKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLWinogradFilterTransformKernel(CLWinogradFilterTransformKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLWinogradFilterTransformKernel &operator=(CLWinogradFilterTransformKernel &&) = default;
-    /** Default destructor */
-    ~CLWinogradFilterTransformKernel() = default;
-    /** Set the input and output tensor.
-     *
-     * @note Winograd filter transform supports the following configurations for NCWH data layout
-     *       F(output tile, kernel size):F(2x2, 3x3), F(2x1, 3x1), F(1x2, 1x3),
-     *                                   F(4x4, 3x3), F(4x1, 3x1), F(1x4, 1x3),
-     *                                   F(4x4, 5x5), F(4x1, 5x1), F(1x4, 1x5)
-     *
-     * @note Winograd filter transform supports the following configurations for NHWC data layout
-     *       F(output tile, kernel size):F(4x4, 3x3), F(4x1, 3x1), F(1x4, 1x3),
-     *                                   F(4x4, 5x5), F(4x1, 5x1), F(1x4, 1x5)
-     *
-     *       Strides: only unit strides
-     *
-     * @param[in]  input         Source tensor. The input is a 4D tensor with dimensions [kernel_x, kernel_y, IFM, OFM] (NCHW data layout) or [IFM, kernel_x, kernel_y, OFM] (NHWC data layout). Data types supported: F16/F32.
-     * @param[out] output        The output tensor. The shape for this tensor can be calculated using the utility function @p compute_winograd_filter_transform_shape. Data types supported: Same as @p input
-     * @param[in]  winograd_info Contains Winograd's information described in @ref WinogradInfo
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, const WinogradInfo &winograd_info);
-    /** Set the input and output tensor.
-     *
-     * @note Winograd filter transform supports the following configurations for NCWH data layout
-     *       F(output tile, kernel size):F(2x2, 3x3), F(2x1, 3x1), F(1x2, 1x3),
-     *                                   F(4x4, 3x3), F(4x1, 3x1), F(1x4, 1x3),
-     *                                   F(4x4, 5x5), F(4x1, 5x1), F(1x4, 1x5)
-     *
-     * @note Winograd filter transform supports the following configurations for NHWC data layout
-     *       F(output tile, kernel size):F(4x4, 3x3), F(4x1, 3x1), F(1x4, 1x3),
-     *                                   F(4x4, 5x5), F(4x1, 5x1), F(1x4, 1x5)
-     *
-     *       Strides: only unit strides
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Source tensor. The input is a 4D tensor with dimensions [kernel_x, kernel_y, IFM, OFM] (NCHW data layout) or [IFM, kernel_x, kernel_y, OFM] (NHWC data layout). Data types supported: F16/F32.
-     * @param[out] output          The output tensor. The shape for this tensor can be calculated using the utility function @p compute_winograd_filter_transform_shape. Data types supported: Same as @p input
-     * @param[in]  winograd_info   Contains Winograd's information described in @ref WinogradInfo
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, const WinogradInfo &winograd_info);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLWinogradFilterTransformKernel
-     *
-     * @note Winograd filter transform supports the following configurations for NCWH data layout
-     *       F(output tile, kernel size):F(2x2, 3x3), F(2x1, 3x1), F(1x2, 1x3),
-     *                                   F(4x4, 3x3), F(4x1, 3x1), F(1x4, 1x3),
-     *                                   F(4x4, 5x5), F(4x1, 5x1), F(1x4, 1x5)
-     *
-     * @note Winograd filter transform supports the following configurations for NHWC data layout
-     *       F(output tile, kernel size):F(4x4, 3x3), F(4x1, 3x1), F(1x4, 1x3),
-     *                                   F(4x4, 5x5), F(4x1, 5x1), F(1x4, 1x5)
-     *
-     *       Strides: only unit strides
-     *
-     * @param[in]  input         Source tensor. The input is a 4D tensor with dimensions [kernel_x, kernel_y, IFM, OFM] (NCHW data layout) or [IFM, kernel_x, kernel_y, OFM] (NHWC data layout). Data types supported: F16/F32.
-     * @param[out] output        The output tensor. The shape for this tensor can be calculated using the utility function @p compute_winograd_filter_transform_shape. Data types supported: Same as @p input
-     * @param[in]  winograd_info Contains Winograd's information described in @ref WinogradInfo
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output, const WinogradInfo &winograd_info);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;
-    ICLTensor       *_output;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLWINOGRADFILTERTRANSFORMKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLWinogradInputTransformKernel.h b/arm_compute/core/CL/kernels/CLWinogradInputTransformKernel.h
deleted file mode 100644
index 4f198f034a..0000000000
--- a/arm_compute/core/CL/kernels/CLWinogradInputTransformKernel.h
+++ /dev/null
@@ -1,121 +0,0 @@
-/*
- * Copyright (c) 2018-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLWINOGRADINPUTTRANSFORMKERNEL_H
-#define ARM_COMPUTE_CLWINOGRADINPUTTRANSFORMKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** OpenCL kernel to perform Winograd input transform.*/
-class CLWinogradInputTransformKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLWinogradInputTransformKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLWinogradInputTransformKernel(const CLWinogradInputTransformKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLWinogradInputTransformKernel &operator=(const CLWinogradInputTransformKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLWinogradInputTransformKernel(CLWinogradInputTransformKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLWinogradInputTransformKernel &operator=(CLWinogradInputTransformKernel &&) = default;
-    /** Set the input and output of the kernel.
-     *
-     * @note Winograd input transform supports the following configurations for NCWH data layout
-     *       F(output tile, kernel size):F(2x2, 3x3), F(2x1, 3x1), F(1x2, 1x3),
-     *                                   F(4x4, 3x3), F(4x1, 3x1), F(1x4, 1x3),
-     *                                   F(4x4, 5x5), F(4x1, 5x1), F(1x4, 1x5)
-     *
-     * @note Winograd input transform supports the following configurations for NHWC data layout
-     *       F(output tile, kernel size):F(4x4, 3x3), F(4x1, 3x1), F(1x4, 1x3),
-     *                                   F(4x4, 5x5), F(4x1, 5x1), F(1x4, 1x5)
-     *
-     *       Strides: only unit strides
-     *
-     * @param[in] input         The input tensor to transform. Data types supported: F16/F32
-     * @param[in] output        The output tensor. The shape for this tensor can be calculated using the utility function @p compute_winograd_input_transform_shape. Data types supported: Same as @p input
-     * @param[in] winograd_info Contains Winograd's information described in @ref WinogradInfo.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, const WinogradInfo &winograd_info);
-    /** Set the input and output of the kernel.
-     *
-     * @note Winograd input transform supports the following configurations for NCWH data layout
-     *       F(output tile, kernel size):F(2x2, 3x3), F(2x1, 3x1), F(1x2, 1x3),
-     *                                   F(4x4, 3x3), F(4x1, 3x1), F(1x4, 1x3),
-     *                                   F(4x4, 5x5), F(4x1, 5x1), F(1x4, 1x5)
-     *
-     * @note Winograd input transform supports the following configurations for NHWC data layout
-     *       F(output tile, kernel size):F(4x4, 3x3), F(4x1, 3x1), F(1x4, 1x3),
-     *                                   F(4x4, 5x5), F(4x1, 5x1), F(1x4, 1x5)
-     *
-     *       Strides: only unit strides
-     *
-     * @param[in] compile_context The compile context to be used.
-     * @param[in] input           The input tensor to transform. Data types supported: F16/F32
-     * @param[in] output          The output tensor. The shape for this tensor can be calculated using the utility function @p compute_winograd_input_transform_shape. Data types supported: Same as @p input
-     * @param[in] winograd_info   Contains Winograd's information described in @ref WinogradInfo.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, const WinogradInfo &winograd_info);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLWinogradInputTransformKernel
-     *
-     * @note Winograd input transform supports the following configurations for NCWH data layout
-     *       F(output tile, kernel size):F(2x2, 3x3), F(2x1, 3x1), F(1x2, 1x3),
-     *                                   F(4x4, 3x3), F(4x1, 3x1), F(1x4, 1x3),
-     *                                   F(4x4, 5x5), F(4x1, 5x1), F(1x4, 1x5)
-     *
-     * @note Winograd input transform supports the following configurations for NHWC data layout
-     *       F(output tile, kernel size):F(4x4, 3x3), F(4x1, 3x1), F(1x4, 1x3),
-     *                                   F(4x4, 5x5), F(4x1, 5x1), F(1x4, 1x5)
-     *
-     *       Strides: only unit strides
-     *
-     * @param[in] input         The input tensor to transform. Data types supported: F16/F32
-     * @param[in] output        The output tensor. The shape for this tensor can be calculated using the utility function @p compute_winograd_input_transform_shape. Data types supported: Same as @p input
-     * @param[in] winograd_info Contains Winograd's information described in @ref WinogradInfo.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output, const WinogradInfo &winograd_info);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-    BorderSize border_size() const override;
-
-private:
-    using WinogradKey = std::pair<std::pair<int, int>, std::pair<int, int>>;
-
-    BorderSize       _border_size;
-    const ICLTensor *_input;
-    ICLTensor       *_output;
-    DataLayout       _data_layout;
-    int              _num_tiles_x;
-    int              _num_tiles_y;
-    unsigned int     _step_z;
-};
-} // arm_compute
-#endif /*ARM_COMPUTE_CLWINOGRADINPUTTRANSFORMKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLWinogradOutputTransformKernel.h b/arm_compute/core/CL/kernels/CLWinogradOutputTransformKernel.h
deleted file mode 100644
index f7cbd05020..0000000000
--- a/arm_compute/core/CL/kernels/CLWinogradOutputTransformKernel.h
+++ /dev/null
@@ -1,127 +0,0 @@
-/*
- * Copyright (c) 2018-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLWINOGRADOUTPUTTRANSFORMKERNEL_H
-#define ARM_COMPUTE_CLWINOGRADOUTPUTTRANSFORMKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the Winograd output transform kernel. */
-class CLWinogradOutputTransformKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLWinogradOutputTransformKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLWinogradOutputTransformKernel(const CLWinogradOutputTransformKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLWinogradOutputTransformKernel &operator=(const CLWinogradOutputTransformKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLWinogradOutputTransformKernel(CLWinogradOutputTransformKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLWinogradOutputTransformKernel &operator=(CLWinogradOutputTransformKernel &&) = default;
-    /** Default destructor */
-    ~CLWinogradOutputTransformKernel() = default;
-    /** Set the input and output tensor.
-     *
-     * @note Winograd output transform supports the following configurations for NCWH data layout
-     *       F(output tile, kernel size):F(2x2, 3x3), F(2x1, 3x1), F(1x2, 1x3),
-     *                                   F(4x4, 3x3), F(4x1, 3x1), F(1x4, 1x3),
-     *                                   F(4x4, 5x5), F(4x1, 5x1), F(1x4, 1x5)
-     *
-     * @note Winograd output transform supports the following configurations for NHWC data layout
-     *       F(output tile, kernel size):F(4x4, 3x3), F(4x1, 3x1), F(1x4, 1x3),
-     *                                   F(4x4, 5x5), F(4x1, 5x1), F(1x4, 1x5)
-     *
-     *       Strides: only unit strides
-     *
-     * @param[in]  input         Source tensor with shape [C, N, K, batches]. Data types supported: F16/F32.
-     * @param[in]  bias          Biases tensor. Shared biases supported. Biases are 1D tensor with dimensions [OFM]. It can be a nullptr. Data type supported: as @p input
-     * @param[out] output        The output tensor. The shape for this tensor can be calculated using the utility function @p compute_winograd_output_transform_shape. Data types supported: Same as @p input
-     * @param[in]  winograd_info Contains Winograd's information described in @ref WinogradInfo
-     * @param[in]  act_info      (Optional) Activation layer information in case of a fused activation.
-     */
-    void configure(const ICLTensor *input, const ICLTensor *bias, ICLTensor *output, const WinogradInfo &winograd_info, const ActivationLayerInfo &act_info = ActivationLayerInfo());
-    /** Set the input and output tensor.
-     *
-     * @note Winograd output transform supports the following configurations for NCWH data layout
-     *       F(output tile, kernel size):F(2x2, 3x3), F(2x1, 3x1), F(1x2, 1x3),
-     *                                   F(4x4, 3x3), F(4x1, 3x1), F(1x4, 1x3),
-     *                                   F(4x4, 5x5), F(4x1, 5x1), F(1x4, 1x5)
-     *
-     * @note Winograd output transform supports the following configurations for NHWC data layout
-     *       F(output tile, kernel size):F(4x4, 3x3), F(4x1, 3x1), F(1x4, 1x3),
-     *                                   F(4x4, 5x5), F(4x1, 5x1), F(1x4, 1x5)
-     *
-     *       Strides: only unit strides
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Source tensor with shape [C, N, K, batches]. Data types supported: F16/F32.
-     * @param[in]  bias            Biases tensor. Shared biases supported. Biases are 1D tensor with dimensions [OFM]. It can be a nullptr. Data type supported: as @p input
-     * @param[out] output          The output tensor. The shape for this tensor can be calculated using the utility function @p compute_winograd_output_transform_shape. Data types supported: Same as @p input
-     * @param[in]  winograd_info   Contains Winograd's information described in @ref WinogradInfo
-     * @param[in]  act_info        (Optional) Activation layer information in case of a fused activation.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, const ICLTensor *bias, ICLTensor *output, const WinogradInfo &winograd_info,
-                   const ActivationLayerInfo &act_info = ActivationLayerInfo());
-
-    /** Static function to check if given info will lead to a valid configuration of @ref CLWinogradOutputTransformKernel
-     *
-     * @note Winograd output transform supports the following configurations for NCWH data layout
-     *       F(output tile, kernel size):F(2x2, 3x3), F(2x1, 3x1), F(1x2, 1x3),
-     *                                   F(4x4, 3x3), F(4x1, 3x1), F(1x4, 1x3),
-     *                                   F(4x4, 5x5), F(4x1, 5x1), F(1x4, 1x5)
-     *
-     * @note Winograd output transform supports the following configurations for NHWC data layout
-     *       F(output tile, kernel size):F(4x4, 3x3), F(4x1, 3x1), F(1x4, 1x3),
-     *                                   F(4x4, 5x5), F(4x1, 5x1), F(1x4, 1x5)
-     *
-     *       Strides: only unit strides
-     *
-     * @param[in]  input         Source tensor with shape [C, N, K, batches]. Data types supported: F16/F32.
-     * @param[in]  bias          Biases tensor. Shared biases supported. Biases are 1D tensor with dimensions [OFM]. It can be a nullptr. Data type supported: as @p input
-     * @param[out] output        The output tensor. The shape for this tensor can be calculated using the utility function @p compute_winograd_output_transform_shape. Data types supported: Same as @p input
-     * @param[in]  winograd_info Contains Winograd's information described in @ref WinogradInfo
-     * @param[in]  act_info      (Optional) Activation layer information in case of a fused activation @ref ActivationLayerInfo. Only RELU, BOUNDED_RELU, LU_BOUNDED_RELU, LEAKY_RELU and SOFT_RELU supported.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *bias, const ITensorInfo *output, const WinogradInfo &winograd_info, const ActivationLayerInfo &act_info = ActivationLayerInfo());
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    using WinogradKey = std::pair<std::pair<int, int>, std::pair<int, int>>;
-
-    const ICLTensor *_input;
-    const ICLTensor *_bias;
-    ICLTensor       *_output;
-    bool             _is_nhwc;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLWINOGRADOUTPUTTRANSFORMKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/CLYOLOLayerKernel.h b/arm_compute/core/CL/kernels/CLYOLOLayerKernel.h
deleted file mode 100644
index 52b069868e..0000000000
--- a/arm_compute/core/CL/kernels/CLYOLOLayerKernel.h
+++ /dev/null
@@ -1,98 +0,0 @@
-/*
- * Copyright (c) 2018-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLYOLOLAYERKERNEL_H
-#define ARM_COMPUTE_CLYOLOLAYERKERNEL_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the YOLO layer kernel that performs partial activation.
- *  For each box, activate only:
- *    - x and y position (channel 0 and 1 of each box)
- *    - objectiveness    (channel 4 of each box)
- *    - classes          (channel 5 to (classes - 5) of each box)
- */
-class CLYOLOLayerKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLYOLOLayerKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLYOLOLayerKernel(const CLYOLOLayerKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLYOLOLayerKernel &operator=(const CLYOLOLayerKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLYOLOLayerKernel(CLYOLOLayerKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLYOLOLayerKernel &operator=(CLYOLOLayerKernel &&) = default;
-    /** Default destructor */
-    ~CLYOLOLayerKernel() = default;
-    /** Set the input and output tensor.
-     *
-     * @note If the output tensor is a nullptr, the activation function will be performed in-place
-     *
-     * @param[in, out] input       Source tensor. In case of @p output tensor = nullptr, this tensor will store the result
-     *                             of the activation function. Data types supported: F16/F32.
-     * @param[out]     output      Destination tensor. Data type supported: same as @p input
-     * @param[in]      act_info    Activation layer information.
-     * @param[in]      num_classes Number of classes to activate (must be submultiple of @p input channels)
-     */
-    void configure(ICLTensor *input, ICLTensor *output, const ActivationLayerInfo &act_info, int32_t num_classes);
-    /** Set the input and output tensor.
-     *
-     * @note If the output tensor is a nullptr, the activation function will be performed in-place
-     *
-     * @param[in]      compile_context The compile context to be used.
-     * @param[in, out] input           Source tensor. In case of @p output tensor = nullptr, this tensor will store the result
-     *                                 of the activation function. Data types supported: F16/F32.
-     * @param[out]     output          Destination tensor. Data type supported: same as @p input
-     * @param[in]      act_info        Activation layer information.
-     * @param[in]      num_classes     Number of classes to activate (must be submultiple of @p input channels)
-     */
-    void configure(const CLCompileContext &compile_context, ICLTensor *input, ICLTensor *output, const ActivationLayerInfo &act_info, int32_t num_classes);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLYOLOLayerKernel
-     *
-     * @param[in] input       Source tensor info. In case of @p output tensor info = nullptr, this tensor will store the result
-     *                        of the activation function. Data types supported: F16/F32.
-     * @param[in] output      Destination tensor info. Data type supported: same as @p input
-     * @param[in] act_info    Activation layer information.
-     * @param[in] num_classes Number of classes to activate (must be submultiple of @p input channels)
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output, const ActivationLayerInfo &act_info, int32_t num_classes);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    ICLTensor *_input;
-    ICLTensor *_output;
-    bool       _run_in_place;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLYOLOLAYERKERNEL_H */
diff --git a/arm_compute/core/CL/kernels/ICLDepthwiseConvolutionLayer3x3Kernel.h b/arm_compute/core/CL/kernels/ICLDepthwiseConvolutionLayer3x3Kernel.h
deleted file mode 100644
index d182e386b8..0000000000
--- a/arm_compute/core/CL/kernels/ICLDepthwiseConvolutionLayer3x3Kernel.h
+++ /dev/null
@@ -1,105 +0,0 @@
-/*
- * Copyright (c) 2017-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_ICLDEPTHWISECONVOLUTIONKERNEL3x3_H
-#define ARM_COMPUTE_ICLDEPTHWISECONVOLUTIONKERNEL3x3_H
-
-#include "arm_compute/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Interface for the kernel to run a 3x3 depthwise convolution on a tensor.
- */
-class ICLDepthwiseConvolutionLayer3x3Kernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    ICLDepthwiseConvolutionLayer3x3Kernel()
-        : _border_size(0), _input(), _output(), _weights(), _biases(), _conv_stride_y(1), _output_multipliers(), _output_shifts(), _is_quantized(false)
-    {
-    }
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    ICLDepthwiseConvolutionLayer3x3Kernel(const ICLDepthwiseConvolutionLayer3x3Kernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    ICLDepthwiseConvolutionLayer3x3Kernel &operator=(const ICLDepthwiseConvolutionLayer3x3Kernel &) = delete;
-    /** Default Move Constructor. */
-    ICLDepthwiseConvolutionLayer3x3Kernel(ICLDepthwiseConvolutionLayer3x3Kernel &&) = default;
-    /** Default move assignment operator */
-    ICLDepthwiseConvolutionLayer3x3Kernel &operator=(ICLDepthwiseConvolutionLayer3x3Kernel &&) = default;
-    /** Initialize the function's source, destination, conv and border_size.
-     *
-     * @param[in]  input              Source tensor. DataType supported: QASYMM8/F16/F32.
-     * @param[in]  weights            Weights tensor. A 3D tensor with dimensions [3, 3, IFM].
-     *                                Data type supported: Same as @p input, QASYMM8/QSYMM8_PER_CHANNEL when input is QASYMM8.
-     * @param[in]  biases             Biases tensor. A 1D tensor with dimensions [IFM]. Must be nullptr if not needed.
-     *                                Data type supported: Same as @p input, S32 when input is QASYMM8.
-     * @param[out] output             Destination tensor. Data type supported: Same as @p input.
-     * @param[in]  conv_info          Padding and stride information to use for the convolution.
-     * @param[in]  depth_multiplier   (Optional) Multiplier to apply to the input's depth in order to retrieve the output's depth. Defaults to 1.
-     * @param[in]  act_info           (Optional) Activation layer information in case of a fused activation. Only RELU, BOUNDED_RELU and LU_BOUNDED_RELU are supported for QASYMM8.
-     * @param[in]  dilation           (Optional) Dilation, in elements, across x and y. Defaults to (1, 1).
-     * @param[in]  output_multipliers (Optional) Output multipliers tensor for quantized computations. In case of per-channel quantization,
-     *                                the number of multipliers must be equal to the number of filters (IFM). Supported data types: S32
-     * @param[in]  output_shifts      (Optional) Output shifts tensor for quantized computations. In case of per-channel quantization,
-     *                                the number of multipliers must be equal to the number of filters (IFM). Supported data types: S32
-     */
-    virtual void configure(const ICLTensor *input, const ICLTensor *weights, const ICLTensor *biases, ICLTensor *output, const PadStrideInfo &conv_info,
-                           unsigned int depth_multiplier = 1, ActivationLayerInfo act_info = ActivationLayerInfo(), const Size2D &dilation = Size2D(1U, 1U),
-                           const ICLTensor *output_multipliers = nullptr, const ICLTensor *output_shifts = nullptr) = 0;
-    /** Initialize the function's source, destination, conv and border_size.
-     *
-     * @param[in]  compile_context    The compile context to be used.
-     * @param[in]  input              Source tensor. DataType supported: QASYMM8/F16/F32.
-     * @param[in]  weights            Weights tensor. A 3D tensor with dimensions [3, 3, IFM].
-     *                                Data type supported: Same as @p input, QASYMM8/QSYMM8_PER_CHANNEL when input is QASYMM8.
-     * @param[in]  biases             Biases tensor. A 1D tensor with dimensions [IFM]. Must be nullptr if not needed.
-     *                                Data type supported: Same as @p input, S32 when input is QASYMM8.
-     * @param[out] output             Destination tensor. Data type supported: Same as @p input.
-     * @param[in]  conv_info          Padding and stride information to use for the convolution.
-     * @param[in]  depth_multiplier   (Optional) Multiplier to apply to the input's depth in order to retrieve the output's depth. Defaults to 1.
-     * @param[in]  act_info           (Optional) Activation layer information in case of a fused activation. Only RELU, BOUNDED_RELU and LU_BOUNDED_RELU are supported for QASYMM8.
-     * @param[in]  dilation           (Optional) Dilation, in elements, across x and y. Defaults to (1, 1).
-     * @param[in]  output_multipliers (Optional) Output multipliers tensor for quantized computations. In case of per-channel quantization,
-     *                                the number of multipliers must be equal to the number of filters (IFM). Supported data types: S32
-     * @param[in]  output_shifts      (Optional) Output shifts tensor for quantized computations. In case of per-channel quantization,
-     *                                the number of multipliers must be equal to the number of filters (IFM). Supported data types: S32
-     */
-    virtual void configure(const CLCompileContext &compile_context, const ICLTensor *input, const ICLTensor *weights, const ICLTensor *biases, ICLTensor *output, const PadStrideInfo &conv_info,
-                           unsigned int depth_multiplier = 1, ActivationLayerInfo act_info = ActivationLayerInfo(), const Size2D &dilation = Size2D(1U, 1U),
-                           const ICLTensor *output_multipliers = nullptr, const ICLTensor *output_shifts = nullptr) = 0;
-
-protected:
-    BorderSize       _border_size;
-    const ICLTensor *_input;
-    ICLTensor       *_output;
-    const ICLTensor *_weights;
-    const ICLTensor *_biases;
-    unsigned int     _conv_stride_y;
-    const ICLTensor *_output_multipliers;
-    const ICLTensor *_output_shifts;
-    bool             _is_quantized;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_ICLDEPTHWISECONVOLUTIONKERNEL3x3_H */