Break up core/Utils.h to reduce unused code being included everywhere

Makes a small difference to compile times and opens up other opportunities
to simplify code.

Change-Id: I232876910bbe4fa9719f4a0ce4a54c090faeb5ef
Signed-off-by: Matthew Bentham <Matthew.Bentham@arm.com>
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/c/VisualCompute/ComputeLibrary/+/532429
Tested-by: bsgcomp <bsgcomp@arm.com>
Reviewed-by: Pablo Tello <pablo.tello@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/9856
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Viet-Hoa Do <viet-hoa.do@arm.com>
Benchmark: Arm Jenkins <bsgcomp@arm.com>

10 files changed, 1115 insertions, 891 deletions

diff --git a/arm_compute/core/TensorInfo.h b/arm_compute/core/TensorInfo.h
index 8436407a75..e738a797b2 100644
--- a/arm_compute/core/TensorInfo.h
+++ b/arm_compute/core/TensorInfo.h
@@ -32,7 +32,6 @@
 #include "arm_compute/core/Strides.h"
 #include "arm_compute/core/TensorShape.h"
 #include "arm_compute/core/Types.h"
-#include "arm_compute/core/Utils.h"
 
 #include <cstddef>
 #include <memory>
diff --git a/arm_compute/core/Utils.h b/arm_compute/core/Utils.h
index 6fa983d24e..c5b50167bf 100644
--- a/arm_compute/core/Utils.h
+++ b/arm_compute/core/Utils.h
@@ -26,21 +26,23 @@
 
 #include "arm_compute/core/Error.h"
 #include "arm_compute/core/PixelValue.h"
-#include "arm_compute/core/Rounding.h"
 #include "arm_compute/core/Types.h"
-#include "arm_compute/core/Version.h"
 
-#include <algorithm>
-#include <cstdint>
-#include <cstdlib>
-#include <iomanip>
+#include <cmath>
 #include <numeric>
 #include <sstream>
 #include <string>
 #include <type_traits>
 #include <unordered_map>
 #include <utility>
-#include <vector>
+
+/* Convenience / backwards compatibility includes */
+#include "arm_compute/core/utils/ActivationFunctionUtils.h"
+#include "arm_compute/core/utils/DataLayoutUtils.h"
+#include "arm_compute/core/utils/DataTypeUtils.h"
+#include "arm_compute/core/utils/FormatUtils.h"
+#include "arm_compute/core/utils/InterpolationPolicyUtils.h"
+#include "arm_compute/core/utils/StringUtils.h"
 
 namespace arm_compute
 {
@@ -57,545 +59,6 @@ class ActivationLayerInfo;
  */
 std::string read_file(const std::string &filename, bool binary);
 
-/** The size in bytes of the data type
- *
- * @param[in] data_type Input data type
- *
- * @return The size in bytes of the data type
- */
-inline size_t data_size_from_type(DataType data_type)
-{
-    switch(data_type)
-    {
-        case DataType::U8:
-        case DataType::S8:
-        case DataType::QSYMM8:
-        case DataType::QASYMM8:
-        case DataType::QASYMM8_SIGNED:
-        case DataType::QSYMM8_PER_CHANNEL:
-            return 1;
-        case DataType::U16:
-        case DataType::S16:
-        case DataType::QSYMM16:
-        case DataType::QASYMM16:
-        case DataType::BFLOAT16:
-        case DataType::F16:
-            return 2;
-        case DataType::F32:
-        case DataType::U32:
-        case DataType::S32:
-            return 4;
-        case DataType::F64:
-        case DataType::U64:
-        case DataType::S64:
-            return 8;
-        case DataType::SIZET:
-            return sizeof(size_t);
-        default:
-            ARM_COMPUTE_ERROR("Invalid data type");
-            return 0;
-    }
-}
-
-/** The size in bytes of the pixel format
- *
- * @param[in] format Input format
- *
- * @return The size in bytes of the pixel format
- */
-inline size_t pixel_size_from_format(Format format)
-{
-    switch(format)
-    {
-        case Format::U8:
-            return 1;
-        case Format::U16:
-        case Format::S16:
-        case Format::BFLOAT16:
-        case Format::F16:
-        case Format::UV88:
-        case Format::YUYV422:
-        case Format::UYVY422:
-            return 2;
-        case Format::RGB888:
-            return 3;
-        case Format::RGBA8888:
-            return 4;
-        case Format::U32:
-        case Format::S32:
-        case Format::F32:
-            return 4;
-        //Doesn't make sense for planar formats:
-        case Format::NV12:
-        case Format::NV21:
-        case Format::IYUV:
-        case Format::YUV444:
-        default:
-            ARM_COMPUTE_ERROR("Undefined pixel size for given format");
-            return 0;
-    }
-}
-
-/** The size in bytes of the data type
- *
- * @param[in] dt Input data type
- *
- * @return The size in bytes of the data type
- */
-inline size_t element_size_from_data_type(DataType dt)
-{
-    switch(dt)
-    {
-        case DataType::S8:
-        case DataType::U8:
-        case DataType::QSYMM8:
-        case DataType::QASYMM8:
-        case DataType::QASYMM8_SIGNED:
-        case DataType::QSYMM8_PER_CHANNEL:
-            return 1;
-        case DataType::U16:
-        case DataType::S16:
-        case DataType::QSYMM16:
-        case DataType::QASYMM16:
-        case DataType::BFLOAT16:
-        case DataType::F16:
-            return 2;
-        case DataType::U32:
-        case DataType::S32:
-        case DataType::F32:
-            return 4;
-        case DataType::U64:
-        case DataType::S64:
-            return 8;
-        default:
-            ARM_COMPUTE_ERROR("Undefined element size for given data type");
-            return 0;
-    }
-}
-
-/** Return the data type used by a given single-planar pixel format
- *
- * @param[in] format Input format
- *
- * @return The size in bytes of the pixel format
- */
-inline DataType data_type_from_format(Format format)
-{
-    switch(format)
-    {
-        case Format::U8:
-        case Format::UV88:
-        case Format::RGB888:
-        case Format::RGBA8888:
-        case Format::YUYV422:
-        case Format::UYVY422:
-            return DataType::U8;
-        case Format::U16:
-            return DataType::U16;
-        case Format::S16:
-            return DataType::S16;
-        case Format::U32:
-            return DataType::U32;
-        case Format::S32:
-            return DataType::S32;
-        case Format::BFLOAT16:
-            return DataType::BFLOAT16;
-        case Format::F16:
-            return DataType::F16;
-        case Format::F32:
-            return DataType::F32;
-        //Doesn't make sense for planar formats:
-        case Format::NV12:
-        case Format::NV21:
-        case Format::IYUV:
-        case Format::YUV444:
-        default:
-            ARM_COMPUTE_ERROR("Not supported data_type for given format");
-            return DataType::UNKNOWN;
-    }
-}
-
-/** Return the plane index of a given channel given an input format.
- *
- * @param[in] format  Input format
- * @param[in] channel Input channel
- *
- * @return The plane index of the specific channel of the specific format
- */
-inline int plane_idx_from_channel(Format format, Channel channel)
-{
-    switch(format)
-    {
-        // Single planar formats have a single plane
-        case Format::U8:
-        case Format::U16:
-        case Format::S16:
-        case Format::U32:
-        case Format::S32:
-        case Format::BFLOAT16:
-        case Format::F16:
-        case Format::F32:
-        case Format::UV88:
-        case Format::RGB888:
-        case Format::RGBA8888:
-        case Format::YUYV422:
-        case Format::UYVY422:
-            return 0;
-        // Multi planar formats
-        case Format::NV12:
-        case Format::NV21:
-        {
-            // Channel U and V share the same plane of format UV88
-            switch(channel)
-            {
-                case Channel::Y:
-                    return 0;
-                case Channel::U:
-                case Channel::V:
-                    return 1;
-                default:
-                    ARM_COMPUTE_ERROR("Not supported channel");
-                    return 0;
-            }
-        }
-        case Format::IYUV:
-        case Format::YUV444:
-        {
-            switch(channel)
-            {
-                case Channel::Y:
-                    return 0;
-                case Channel::U:
-                    return 1;
-                case Channel::V:
-                    return 2;
-                default:
-                    ARM_COMPUTE_ERROR("Not supported channel");
-                    return 0;
-            }
-        }
-        default:
-            ARM_COMPUTE_ERROR("Not supported format");
-            return 0;
-    }
-}
-
-/** Return the channel index of a given channel given an input format.
- *
- * @param[in] format  Input format
- * @param[in] channel Input channel
- *
- * @return The channel index of the specific channel of the specific format
- */
-inline int channel_idx_from_format(Format format, Channel channel)
-{
-    switch(format)
-    {
-        case Format::RGB888:
-        {
-            switch(channel)
-            {
-                case Channel::R:
-                    return 0;
-                case Channel::G:
-                    return 1;
-                case Channel::B:
-                    return 2;
-                default:
-                    ARM_COMPUTE_ERROR("Not supported channel");
-                    return 0;
-            }
-        }
-        case Format::RGBA8888:
-        {
-            switch(channel)
-            {
-                case Channel::R:
-                    return 0;
-                case Channel::G:
-                    return 1;
-                case Channel::B:
-                    return 2;
-                case Channel::A:
-                    return 3;
-                default:
-                    ARM_COMPUTE_ERROR("Not supported channel");
-                    return 0;
-            }
-        }
-        case Format::YUYV422:
-        {
-            switch(channel)
-            {
-                case Channel::Y:
-                    return 0;
-                case Channel::U:
-                    return 1;
-                case Channel::V:
-                    return 3;
-                default:
-                    ARM_COMPUTE_ERROR("Not supported channel");
-                    return 0;
-            }
-        }
-        case Format::UYVY422:
-        {
-            switch(channel)
-            {
-                case Channel::Y:
-                    return 1;
-                case Channel::U:
-                    return 0;
-                case Channel::V:
-                    return 2;
-                default:
-                    ARM_COMPUTE_ERROR("Not supported channel");
-                    return 0;
-            }
-        }
-        case Format::NV12:
-        {
-            switch(channel)
-            {
-                case Channel::Y:
-                    return 0;
-                case Channel::U:
-                    return 0;
-                case Channel::V:
-                    return 1;
-                default:
-                    ARM_COMPUTE_ERROR("Not supported channel");
-                    return 0;
-            }
-        }
-        case Format::NV21:
-        {
-            switch(channel)
-            {
-                case Channel::Y:
-                    return 0;
-                case Channel::U:
-                    return 1;
-                case Channel::V:
-                    return 0;
-                default:
-                    ARM_COMPUTE_ERROR("Not supported channel");
-                    return 0;
-            }
-        }
-        case Format::YUV444:
-        case Format::IYUV:
-        {
-            switch(channel)
-            {
-                case Channel::Y:
-                    return 0;
-                case Channel::U:
-                    return 0;
-                case Channel::V:
-                    return 0;
-                default:
-                    ARM_COMPUTE_ERROR("Not supported channel");
-                    return 0;
-            }
-        }
-        default:
-            ARM_COMPUTE_ERROR("Not supported format");
-            return 0;
-    }
-}
-
-/** Return the number of planes for a given format
- *
- * @param[in] format Input format
- *
- * @return The number of planes for a given image format.
- */
-inline size_t num_planes_from_format(Format format)
-{
-    switch(format)
-    {
-        case Format::U8:
-        case Format::S16:
-        case Format::U16:
-        case Format::S32:
-        case Format::U32:
-        case Format::BFLOAT16:
-        case Format::F16:
-        case Format::F32:
-        case Format::RGB888:
-        case Format::RGBA8888:
-        case Format::YUYV422:
-        case Format::UYVY422:
-            return 1;
-        case Format::NV12:
-        case Format::NV21:
-            return 2;
-        case Format::IYUV:
-        case Format::YUV444:
-            return 3;
-        default:
-            ARM_COMPUTE_ERROR("Not supported format");
-            return 0;
-    }
-}
-
-/** Return the number of channels for a given single-planar pixel format
- *
- * @param[in] format Input format
- *
- * @return The number of channels for a given image format.
- */
-inline size_t num_channels_from_format(Format format)
-{
-    switch(format)
-    {
-        case Format::U8:
-        case Format::U16:
-        case Format::S16:
-        case Format::U32:
-        case Format::S32:
-        case Format::BFLOAT16:
-        case Format::F16:
-        case Format::F32:
-            return 1;
-        // Because the U and V channels are subsampled
-        // these formats appear like having only 2 channels:
-        case Format::YUYV422:
-        case Format::UYVY422:
-            return 2;
-        case Format::UV88:
-            return 2;
-        case Format::RGB888:
-            return 3;
-        case Format::RGBA8888:
-            return 4;
-        //Doesn't make sense for planar formats:
-        case Format::NV12:
-        case Format::NV21:
-        case Format::IYUV:
-        case Format::YUV444:
-        default:
-            return 0;
-    }
-}
-
-/** Return the promoted data type of a given data type.
- *
- * @note If promoted data type is not supported an error will be thrown
- *
- * @param[in] dt Data type to get the promoted type of.
- *
- * @return Promoted data type
- */
-inline DataType get_promoted_data_type(DataType dt)
-{
-    switch(dt)
-    {
-        case DataType::U8:
-            return DataType::U16;
-        case DataType::S8:
-            return DataType::S16;
-        case DataType::U16:
-            return DataType::U32;
-        case DataType::S16:
-            return DataType::S32;
-        case DataType::QSYMM8:
-        case DataType::QASYMM8:
-        case DataType::QASYMM8_SIGNED:
-        case DataType::QSYMM8_PER_CHANNEL:
-        case DataType::QSYMM16:
-        case DataType::QASYMM16:
-        case DataType::BFLOAT16:
-        case DataType::F16:
-        case DataType::U32:
-        case DataType::S32:
-        case DataType::F32:
-            ARM_COMPUTE_ERROR("Unsupported data type promotions!");
-        default:
-            ARM_COMPUTE_ERROR("Undefined data type!");
-    }
-    return DataType::UNKNOWN;
-}
-
-/** Compute the mininum and maximum values a data type can take
- *
- * @param[in] dt Data type to get the min/max bounds of
- *
- * @return A tuple (min,max) with the minimum and maximum values respectively wrapped in PixelValue.
- */
-inline std::tuple<PixelValue, PixelValue> get_min_max(DataType dt)
-{
-    PixelValue min{};
-    PixelValue max{};
-    switch(dt)
-    {
-        case DataType::U8:
-        case DataType::QASYMM8:
-        {
-            min = PixelValue(static_cast<int32_t>(std::numeric_limits<uint8_t>::lowest()));
-            max = PixelValue(static_cast<int32_t>(std::numeric_limits<uint8_t>::max()));
-            break;
-        }
-        case DataType::S8:
-        case DataType::QSYMM8:
-        case DataType::QASYMM8_SIGNED:
-        case DataType::QSYMM8_PER_CHANNEL:
-        {
-            min = PixelValue(static_cast<int32_t>(std::numeric_limits<int8_t>::lowest()));
-            max = PixelValue(static_cast<int32_t>(std::numeric_limits<int8_t>::max()));
-            break;
-        }
-        case DataType::U16:
-        case DataType::QASYMM16:
-        {
-            min = PixelValue(static_cast<int32_t>(std::numeric_limits<uint16_t>::lowest()));
-            max = PixelValue(static_cast<int32_t>(std::numeric_limits<uint16_t>::max()));
-            break;
-        }
-        case DataType::S16:
-        case DataType::QSYMM16:
-        {
-            min = PixelValue(static_cast<int32_t>(std::numeric_limits<int16_t>::lowest()));
-            max = PixelValue(static_cast<int32_t>(std::numeric_limits<int16_t>::max()));
-            break;
-        }
-        case DataType::U32:
-        {
-            min = PixelValue(std::numeric_limits<uint32_t>::lowest());
-            max = PixelValue(std::numeric_limits<uint32_t>::max());
-            break;
-        }
-        case DataType::S32:
-        {
-            min = PixelValue(std::numeric_limits<int32_t>::lowest());
-            max = PixelValue(std::numeric_limits<int32_t>::max());
-            break;
-        }
-        case DataType::BFLOAT16:
-        {
-            min = PixelValue(bfloat16::lowest());
-            max = PixelValue(bfloat16::max());
-            break;
-        }
-        case DataType::F16:
-        {
-            min = PixelValue(std::numeric_limits<half>::lowest());
-            max = PixelValue(std::numeric_limits<half>::max());
-            break;
-        }
-        case DataType::F32:
-        {
-            min = PixelValue(std::numeric_limits<float>::lowest());
-            max = PixelValue(std::numeric_limits<float>::max());
-            break;
-        }
-        default:
-            ARM_COMPUTE_ERROR("Undefined data type!");
-    }
-    return std::make_tuple(min, max);
-}
-
 /** Permutes the given dimensions according the permutation vector
  *
  * @param[in,out] dimensions Dimensions to be permuted.
@@ -717,14 +180,6 @@ QuantizationInfo get_softmax_output_quantization_info(DataType input_type, bool
  */
 std::pair<int32_t, int32_t> get_quantized_activation_min_max(const ActivationLayerInfo &act_info, DataType data_type, UniformQuantizationInfo oq_info);
 
-/** Convert a tensor format into a string.
- *
- * @param[in] format @ref Format to be translated to string.
- *
- * @return The string describing the format.
- */
-const std::string &string_from_format(Format format);
-
 /** Convert a channel identity into a string.
  *
  * @param[in] channel @ref Channel to be translated to string.
@@ -732,34 +187,7 @@ const std::string &string_from_format(Format format);
  * @return The string describing the channel.
  */
 const std::string &string_from_channel(Channel channel);
-/** Convert a data layout identity into a string.
- *
- * @param[in] dl @ref DataLayout to be translated to string.
- *
- * @return The string describing the data layout.
- */
-const std::string &string_from_data_layout(DataLayout dl);
-/** Convert a data type identity into a string.
- *
- * @param[in] dt @ref DataType to be translated to string.
- *
- * @return The string describing the data type.
- */
-const std::string &string_from_data_type(DataType dt);
-/** Translates a given activation function to a string.
- *
- * @param[in] act @ref ActivationLayerInfo::ActivationFunction to be translated to string.
- *
- * @return The string describing the activation function.
- */
-const std::string &string_from_activation_func(const ActivationFunction &act);
-/** Translates a given interpolation policy to a string.
- *
- * @param[in] policy @ref InterpolationPolicy to be translated to string.
- *
- * @return The string describing the interpolation policy.
- */
-const std::string &string_from_interpolation_policy(InterpolationPolicy policy);
+
 /** Translates a given border mode policy to a string.
  *
  * @param[in] border_mode @ref BorderMode to be translated to string.
@@ -820,13 +248,7 @@ const std::string &string_from_gemmlowp_output_stage(GEMMLowpOutputStageType out
  * @return String representation of the PixelValue through the given data type.
  */
 std::string string_from_pixel_value(const PixelValue &value, const DataType data_type);
-/** Convert a string to DataType
- *
- * @param[in] name The name of the data type
- *
- * @return DataType
- */
-DataType data_type_from_name(const std::string &name);
+
 /** Stores padding information before configuring a kernel
  *
  * @param[in] infos list of tensor infos to store the padding info for
@@ -849,170 +271,6 @@ std::unordered_map<const ITensorInfo *, PaddingSize> get_padding_info(std::initi
  */
 bool has_padding_changed(const std::unordered_map<const ITensorInfo *, PaddingSize> &padding_map);
 
-/** Input Stream operator for @ref DataType
- *
- * @param[in]  stream    Stream to parse
- * @param[out] data_type Output data type
- *
- * @return Updated stream
- */
-inline ::std::istream &operator>>(::std::istream &stream, DataType &data_type)
-{
-    std::string value;
-    stream >> value;
-    data_type = data_type_from_name(value);
-    return stream;
-}
-/** Lower a given string.
- *
- * @param[in] val Given string to lower.
- *
- * @return The lowered string
- */
-std::string lower_string(const std::string &val);
-
-/** Raise a given string to upper case
- *
- * @param[in] val Given string to lower.
- *
- * @return The upper case string
- */
-std::string upper_string(const std::string &val);
-
-/** Check if a given data type is of floating point type
- *
- * @param[in] dt Input data type.
- *
- * @return True if data type is of floating point type, else false.
- */
-inline bool is_data_type_float(DataType dt)
-{
-    switch(dt)
-    {
-        case DataType::F16:
-        case DataType::F32:
-            return true;
-        default:
-            return false;
-    }
-}
-
-/** Check if a given data type is of quantized type
- *
- * @note Quantized is considered a super-set of fixed-point and asymmetric data types.
- *
- * @param[in] dt Input data type.
- *
- * @return True if data type is of quantized type, else false.
- */
-inline bool is_data_type_quantized(DataType dt)
-{
-    switch(dt)
-    {
-        case DataType::QSYMM8:
-        case DataType::QASYMM8:
-        case DataType::QASYMM8_SIGNED:
-        case DataType::QSYMM8_PER_CHANNEL:
-        case DataType::QSYMM16:
-        case DataType::QASYMM16:
-            return true;
-        default:
-            return false;
-    }
-}
-
-/** Check if a given data type is of asymmetric quantized type
- *
- * @param[in] dt Input data type.
- *
- * @return True if data type is of asymmetric quantized type, else false.
- */
-inline bool is_data_type_quantized_asymmetric(DataType dt)
-{
-    switch(dt)
-    {
-        case DataType::QASYMM8:
-        case DataType::QASYMM8_SIGNED:
-        case DataType::QASYMM16:
-            return true;
-        default:
-            return false;
-    }
-}
-
-/** Check if a given data type is of asymmetric quantized signed type
- *
- * @param[in] dt Input data type.
- *
- * @return True if data type is of asymmetric quantized signed type, else false.
- */
-inline bool is_data_type_quantized_asymmetric_signed(DataType dt)
-{
-    switch(dt)
-    {
-        case DataType::QASYMM8_SIGNED:
-            return true;
-        default:
-            return false;
-    }
-}
-
-/** Check if a given data type is of symmetric quantized type
- *
- * @param[in] dt Input data type.
- *
- * @return True if data type is of symmetric quantized type, else false.
- */
-inline bool is_data_type_quantized_symmetric(DataType dt)
-{
-    switch(dt)
-    {
-        case DataType::QSYMM8:
-        case DataType::QSYMM8_PER_CHANNEL:
-        case DataType::QSYMM16:
-            return true;
-        default:
-            return false;
-    }
-}
-
-/** Check if a given data type is of per channel type
- *
- * @param[in] dt Input data type.
- *
- * @return True if data type is of per channel type, else false.
- */
-inline bool is_data_type_quantized_per_channel(DataType dt)
-{
-    switch(dt)
-    {
-        case DataType::QSYMM8_PER_CHANNEL:
-            return true;
-        default:
-            return false;
-    }
-}
-
-/** Create a string with the float in full precision.
- *
- * @param val Floating point value
- *
- * @return String with the floating point value.
- */
-inline std::string float_to_string_with_full_precision(float val)
-{
-    std::stringstream ss;
-    ss.precision(std::numeric_limits<float>::max_digits10);
-    ss << val;
-
-    if(val != static_cast<int>(val))
-    {
-        ss << "f";
-    }
-
-    return ss.str();
-}
-
 /** Returns the number of elements required to go from start to end with the wanted step
  *
  * @param[in] start start value
@@ -1027,142 +285,6 @@ inline size_t num_of_elements_in_range(const float start, const float end, const
     return size_t(std::ceil((end - start) / step));
 }
 
-/** Returns true if the value can be represented by the given data type
- *
- * @param[in] val   value to be checked
- * @param[in] dt    data type that is checked
- * @param[in] qinfo (Optional) quantization info if the data type is QASYMM8
- *
- * @return true if the data type can hold the value.
- */
-template <typename T>
-bool check_value_range(T val, DataType dt, QuantizationInfo qinfo = QuantizationInfo())
-{
-    switch(dt)
-    {
-        case DataType::U8:
-        {
-            const auto val_u8 = static_cast<uint8_t>(val);
-            return ((val_u8 == val) && val >= std::numeric_limits<uint8_t>::lowest() && val <= std::numeric_limits<uint8_t>::max());
-        }
-        case DataType::QASYMM8:
-        {
-            double min = static_cast<double>(dequantize_qasymm8(0, qinfo));
-            double max = static_cast<double>(dequantize_qasymm8(std::numeric_limits<uint8_t>::max(), qinfo));
-            return ((double)val >= min && (double)val <= max);
-        }
-        case DataType::S8:
-        {
-            const auto val_s8 = static_cast<int8_t>(val);
-            return ((val_s8 == val) && val >= std::numeric_limits<int8_t>::lowest() && val <= std::numeric_limits<int8_t>::max());
-        }
-        case DataType::U16:
-        {
-            const auto val_u16 = static_cast<uint16_t>(val);
-            return ((val_u16 == val) && val >= std::numeric_limits<uint16_t>::lowest() && val <= std::numeric_limits<uint16_t>::max());
-        }
-        case DataType::S16:
-        {
-            const auto val_s16 = static_cast<int16_t>(val);
-            return ((val_s16 == val) && val >= std::numeric_limits<int16_t>::lowest() && val <= std::numeric_limits<int16_t>::max());
-        }
-        case DataType::U32:
-        {
-            const auto val_d64 = static_cast<double>(val);
-            const auto val_u32 = static_cast<uint32_t>(val);
-            return ((val_u32 == val_d64) && val_d64 >= std::numeric_limits<uint32_t>::lowest() && val_d64 <= std::numeric_limits<uint32_t>::max());
-        }
-        case DataType::S32:
-        {
-            const auto val_d64 = static_cast<double>(val);
-            const auto val_s32 = static_cast<int32_t>(val);
-            return ((val_s32 == val_d64) && val_d64 >= std::numeric_limits<int32_t>::lowest() && val_d64 <= std::numeric_limits<int32_t>::max());
-        }
-        case DataType::BFLOAT16:
-            return (val >= bfloat16::lowest() && val <= bfloat16::max());
-        case DataType::F16:
-            return (val >= std::numeric_limits<half>::lowest() && val <= std::numeric_limits<half>::max());
-        case DataType::F32:
-            return (val >= std::numeric_limits<float>::lowest() && val <= std::numeric_limits<float>::max());
-        default:
-            ARM_COMPUTE_ERROR("Data type not supported");
-            return false;
-    }
-}
-
-/** Returns the adjusted vector size in case it is less than the input's first dimension, getting rounded down to its closest valid vector size
- *
- * @param[in] vec_size vector size to be adjusted
- * @param[in] dim0     size of the first dimension
- *
- * @return the number of element processed along the X axis per thread
- */
-inline unsigned int adjust_vec_size(unsigned int vec_size, size_t dim0)
-{
-    ARM_COMPUTE_ERROR_ON(vec_size > 16);
-
-    if((vec_size >= dim0) && (dim0 == 3))
-    {
-        return dim0;
-    }
-
-    while(vec_size > dim0)
-    {
-        vec_size >>= 1;
-    }
-
-    return vec_size;
-}
-
-/** Returns the suffix string of CPU kernel implementation names based on the given data type
- *
- * @param[in] data_type The data type the CPU kernel implemetation uses
- *
- * @return the suffix string of CPU kernel implementations
- */
-inline std::string cpu_impl_dt(const DataType &data_type)
-{
-    std::string ret = "";
-
-    switch(data_type)
-    {
-        case DataType::F32:
-            ret = "fp32";
-            break;
-        case DataType::F16:
-            ret = "fp16";
-            break;
-        case DataType::U8:
-            ret = "u8";
-            break;
-        case DataType::S16:
-            ret = "s16";
-            break;
-        case DataType::S32:
-            ret = "s32";
-            break;
-        case DataType::QASYMM8:
-            ret = "qu8";
-            break;
-        case DataType::QASYMM8_SIGNED:
-            ret = "qs8";
-            break;
-        case DataType::QSYMM16:
-            ret = "qs16";
-            break;
-        case DataType::QSYMM8_PER_CHANNEL:
-            ret = "qp8";
-            break;
-        case DataType::BFLOAT16:
-            ret = "bf16";
-            break;
-        default:
-            ARM_COMPUTE_ERROR("Unsupported.");
-    }
-
-    return ret;
-}
-
 #ifdef ARM_COMPUTE_ASSERTS_ENABLED
 /** Print consecutive elements to an output stream.
  *
diff --git a/arm_compute/core/Validate.h b/arm_compute/core/Validate.h
index ff4540cd8a..cf7eb8b663 100644
--- a/arm_compute/core/Validate.h
+++ b/arm_compute/core/Validate.h
@@ -24,11 +24,13 @@
 #ifndef ARM_COMPUTE_VALIDATE_H
 #define ARM_COMPUTE_VALIDATE_H
 
+#include "arm_compute/core/utils/DataLayoutUtils.h"
+#include "arm_compute/core/utils/DataTypeUtils.h"
 #include "arm_compute/core/Error.h"
+#include "arm_compute/core/utils/FormatUtils.h"
 #include "arm_compute/core/IKernel.h"
 #include "arm_compute/core/ITensor.h"
 #include "arm_compute/core/QuantizationInfo.h"
-#include "arm_compute/core/Utils.h"
 #include "arm_compute/core/Window.h"
 
 #include <algorithm>
diff --git a/arm_compute/core/utils/ActivationFunctionUtils.h b/arm_compute/core/utils/ActivationFunctionUtils.h
new file mode 100644
index 0000000000..1cb66da13d
--- /dev/null
+++ b/arm_compute/core/utils/ActivationFunctionUtils.h
@@ -0,0 +1,41 @@
+/*
+ * Copyright (c) 2016-2023 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_CORE_UTILS_ACTIVATIONFUNCTIONUTILS_H
+#define ARM_COMPUTE_CORE_UTILS_ACTIVATIONFUNCTIONUTILS_H
+
+#include "arm_compute/core/Types.h"
+
+#include <string>
+
+namespace arm_compute
+{
+/** Translates a given activation function to a string.
+ *
+ * @param[in] act @ref ActivationLayerInfo::ActivationFunction to be translated to string.
+ *
+ * @return The string describing the activation function.
+ */
+const std::string &string_from_activation_func(const ActivationFunction &act);
+}
+#endif /*ARM_COMPUTE_CORE_UTILS_ACTIVATIONFUNCTIONUTILS_H */
diff --git a/arm_compute/core/utils/DataLayoutUtils.h b/arm_compute/core/utils/DataLayoutUtils.h
new file mode 100644
index 0000000000..399f55c63f
--- /dev/null
+++ b/arm_compute/core/utils/DataLayoutUtils.h
@@ -0,0 +1,40 @@
+/*
+ * Copyright (c) 2016-2023 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_CORE_UTILS_DATALAYOUTUTILS_H
+#define ARM_COMPUTE_CORE_UTILS_DATALAYOUTUTILS_H
+#include "arm_compute/core/Types.h"
+
+#include <string>
+
+namespace arm_compute
+{
+/** Convert a data layout identity into a string.
+ *
+ * @param[in] dl @ref DataLayout to be translated to string.
+ *
+ * @return The string describing the data layout.
+ */
+const std::string &string_from_data_layout(DataLayout dl);
+}
+#endif /*ARM_COMPUTE_CORE_UTILS_DATALAYOUTUTILS_H */
diff --git a/arm_compute/core/utils/DataTypeUtils.h b/arm_compute/core/utils/DataTypeUtils.h
new file mode 100644
index 0000000000..cbb409c8a1
--- /dev/null
+++ b/arm_compute/core/utils/DataTypeUtils.h
@@ -0,0 +1,525 @@
+/*
+ * Copyright (c) 2016-2023 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_CORE_UTILS_DATATYPEUTILS_H
+#define ARM_COMPUTE_CORE_UTILS_DATATYPEUTILS_H
+
+#include "arm_compute/core/PixelValue.h"
+#include "arm_compute/core/Types.h"
+
+namespace arm_compute
+{
+/** The size in bytes of the data type
+ *
+ * @param[in] data_type Input data type
+ *
+ * @return The size in bytes of the data type
+ */
+inline size_t data_size_from_type(DataType data_type)
+{
+    switch(data_type)
+    {
+        case DataType::U8:
+        case DataType::S8:
+        case DataType::QSYMM8:
+        case DataType::QASYMM8:
+        case DataType::QASYMM8_SIGNED:
+        case DataType::QSYMM8_PER_CHANNEL:
+            return 1;
+        case DataType::U16:
+        case DataType::S16:
+        case DataType::QSYMM16:
+        case DataType::QASYMM16:
+        case DataType::BFLOAT16:
+        case DataType::F16:
+            return 2;
+        case DataType::F32:
+        case DataType::U32:
+        case DataType::S32:
+            return 4;
+        case DataType::F64:
+        case DataType::U64:
+        case DataType::S64:
+            return 8;
+        case DataType::SIZET:
+            return sizeof(size_t);
+        default:
+            ARM_COMPUTE_ERROR("Invalid data type");
+            return 0;
+    }
+}
+
+/** The size in bytes of the data type
+ *
+ * @param[in] dt Input data type
+ *
+ * @return The size in bytes of the data type
+ */
+inline size_t element_size_from_data_type(DataType dt)
+{
+    switch(dt)
+    {
+        case DataType::S8:
+        case DataType::U8:
+        case DataType::QSYMM8:
+        case DataType::QASYMM8:
+        case DataType::QASYMM8_SIGNED:
+        case DataType::QSYMM8_PER_CHANNEL:
+            return 1;
+        case DataType::U16:
+        case DataType::S16:
+        case DataType::QSYMM16:
+        case DataType::QASYMM16:
+        case DataType::BFLOAT16:
+        case DataType::F16:
+            return 2;
+        case DataType::U32:
+        case DataType::S32:
+        case DataType::F32:
+            return 4;
+        case DataType::U64:
+        case DataType::S64:
+            return 8;
+        default:
+            ARM_COMPUTE_ERROR("Undefined element size for given data type");
+            return 0;
+    }
+}
+
+/** Return the data type used by a given single-planar pixel format
+ *
+ * @param[in] format Input format
+ *
+ * @return The size in bytes of the pixel format
+ */
+inline DataType data_type_from_format(Format format)
+{
+    switch(format)
+    {
+        case Format::U8:
+        case Format::UV88:
+        case Format::RGB888:
+        case Format::RGBA8888:
+        case Format::YUYV422:
+        case Format::UYVY422:
+            return DataType::U8;
+        case Format::U16:
+            return DataType::U16;
+        case Format::S16:
+            return DataType::S16;
+        case Format::U32:
+            return DataType::U32;
+        case Format::S32:
+            return DataType::S32;
+        case Format::BFLOAT16:
+            return DataType::BFLOAT16;
+        case Format::F16:
+            return DataType::F16;
+        case Format::F32:
+            return DataType::F32;
+        //Doesn't make sense for planar formats:
+        case Format::NV12:
+        case Format::NV21:
+        case Format::IYUV:
+        case Format::YUV444:
+        default:
+            ARM_COMPUTE_ERROR("Not supported data_type for given format");
+            return DataType::UNKNOWN;
+    }
+}
+
+/** Return the promoted data type of a given data type.
+ *
+ * @note If promoted data type is not supported an error will be thrown
+ *
+ * @param[in] dt Data type to get the promoted type of.
+ *
+ * @return Promoted data type
+ */
+inline DataType get_promoted_data_type(DataType dt)
+{
+    switch(dt)
+    {
+        case DataType::U8:
+            return DataType::U16;
+        case DataType::S8:
+            return DataType::S16;
+        case DataType::U16:
+            return DataType::U32;
+        case DataType::S16:
+            return DataType::S32;
+        case DataType::QSYMM8:
+        case DataType::QASYMM8:
+        case DataType::QASYMM8_SIGNED:
+        case DataType::QSYMM8_PER_CHANNEL:
+        case DataType::QSYMM16:
+        case DataType::QASYMM16:
+        case DataType::BFLOAT16:
+        case DataType::F16:
+        case DataType::U32:
+        case DataType::S32:
+        case DataType::F32:
+            ARM_COMPUTE_ERROR("Unsupported data type promotions!");
+        default:
+            ARM_COMPUTE_ERROR("Undefined data type!");
+    }
+    return DataType::UNKNOWN;
+}
+
+/** Compute the mininum and maximum values a data type can take
+ *
+ * @param[in] dt Data type to get the min/max bounds of
+ *
+ * @return A tuple (min,max) with the minimum and maximum values respectively wrapped in PixelValue.
+ */
+inline std::tuple<PixelValue, PixelValue> get_min_max(DataType dt)
+{
+    PixelValue min{};
+    PixelValue max{};
+    switch(dt)
+    {
+        case DataType::U8:
+        case DataType::QASYMM8:
+        {
+            min = PixelValue(static_cast<int32_t>(std::numeric_limits<uint8_t>::lowest()));
+            max = PixelValue(static_cast<int32_t>(std::numeric_limits<uint8_t>::max()));
+            break;
+        }
+        case DataType::S8:
+        case DataType::QSYMM8:
+        case DataType::QASYMM8_SIGNED:
+        case DataType::QSYMM8_PER_CHANNEL:
+        {
+            min = PixelValue(static_cast<int32_t>(std::numeric_limits<int8_t>::lowest()));
+            max = PixelValue(static_cast<int32_t>(std::numeric_limits<int8_t>::max()));
+            break;
+        }
+        case DataType::U16:
+        case DataType::QASYMM16:
+        {
+            min = PixelValue(static_cast<int32_t>(std::numeric_limits<uint16_t>::lowest()));
+            max = PixelValue(static_cast<int32_t>(std::numeric_limits<uint16_t>::max()));
+            break;
+        }
+        case DataType::S16:
+        case DataType::QSYMM16:
+        {
+            min = PixelValue(static_cast<int32_t>(std::numeric_limits<int16_t>::lowest()));
+            max = PixelValue(static_cast<int32_t>(std::numeric_limits<int16_t>::max()));
+            break;
+        }
+        case DataType::U32:
+        {
+            min = PixelValue(std::numeric_limits<uint32_t>::lowest());
+            max = PixelValue(std::numeric_limits<uint32_t>::max());
+            break;
+        }
+        case DataType::S32:
+        {
+            min = PixelValue(std::numeric_limits<int32_t>::lowest());
+            max = PixelValue(std::numeric_limits<int32_t>::max());
+            break;
+        }
+        case DataType::BFLOAT16:
+        {
+            min = PixelValue(bfloat16::lowest());
+            max = PixelValue(bfloat16::max());
+            break;
+        }
+        case DataType::F16:
+        {
+            min = PixelValue(std::numeric_limits<half>::lowest());
+            max = PixelValue(std::numeric_limits<half>::max());
+            break;
+        }
+        case DataType::F32:
+        {
+            min = PixelValue(std::numeric_limits<float>::lowest());
+            max = PixelValue(std::numeric_limits<float>::max());
+            break;
+        }
+        default:
+            ARM_COMPUTE_ERROR("Undefined data type!");
+    }
+    return std::make_tuple(min, max);
+}
+
+/** Convert a data type identity into a string.
+ *
+ * @param[in] dt @ref DataType to be translated to string.
+ *
+ * @return The string describing the data type.
+ */
+const std::string &string_from_data_type(DataType dt);
+
+/** Convert a string to DataType
+ *
+ * @param[in] name The name of the data type
+ *
+ * @return DataType
+ */
+DataType data_type_from_name(const std::string &name);
+
+/** Input Stream operator for @ref DataType
+ *
+ * @param[in]  stream    Stream to parse
+ * @param[out] data_type Output data type
+ *
+ * @return Updated stream
+ */
+inline ::std::istream &operator>>(::std::istream &stream, DataType &data_type)
+{
+    std::string value;
+    stream >> value;
+    data_type = data_type_from_name(value);
+    return stream;
+}
+
+/** Check if a given data type is of floating point type
+ *
+ * @param[in] dt Input data type.
+ *
+ * @return True if data type is of floating point type, else false.
+ */
+inline bool is_data_type_float(DataType dt)
+{
+    switch(dt)
+    {
+        case DataType::F16:
+        case DataType::F32:
+            return true;
+        default:
+            return false;
+    }
+}
+
+/** Check if a given data type is of quantized type
+ *
+ * @note Quantized is considered a super-set of fixed-point and asymmetric data types.
+ *
+ * @param[in] dt Input data type.
+ *
+ * @return True if data type is of quantized type, else false.
+ */
+inline bool is_data_type_quantized(DataType dt)
+{
+    switch(dt)
+    {
+        case DataType::QSYMM8:
+        case DataType::QASYMM8:
+        case DataType::QASYMM8_SIGNED:
+        case DataType::QSYMM8_PER_CHANNEL:
+        case DataType::QSYMM16:
+        case DataType::QASYMM16:
+            return true;
+        default:
+            return false;
+    }
+}
+
+/** Check if a given data type is of asymmetric quantized type
+ *
+ * @param[in] dt Input data type.
+ *
+ * @return True if data type is of asymmetric quantized type, else false.
+ */
+inline bool is_data_type_quantized_asymmetric(DataType dt)
+{
+    switch(dt)
+    {
+        case DataType::QASYMM8:
+        case DataType::QASYMM8_SIGNED:
+        case DataType::QASYMM16:
+            return true;
+        default:
+            return false;
+    }
+}
+
+/** Check if a given data type is of asymmetric quantized signed type
+ *
+ * @param[in] dt Input data type.
+ *
+ * @return True if data type is of asymmetric quantized signed type, else false.
+ */
+inline bool is_data_type_quantized_asymmetric_signed(DataType dt)
+{
+    switch(dt)
+    {
+        case DataType::QASYMM8_SIGNED:
+            return true;
+        default:
+            return false;
+    }
+}
+
+/** Check if a given data type is of symmetric quantized type
+ *
+ * @param[in] dt Input data type.
+ *
+ * @return True if data type is of symmetric quantized type, else false.
+ */
+inline bool is_data_type_quantized_symmetric(DataType dt)
+{
+    switch(dt)
+    {
+        case DataType::QSYMM8:
+        case DataType::QSYMM8_PER_CHANNEL:
+        case DataType::QSYMM16:
+            return true;
+        default:
+            return false;
+    }
+}
+
+/** Check if a given data type is of per channel type
+ *
+ * @param[in] dt Input data type.
+ *
+ * @return True if data type is of per channel type, else false.
+ */
+inline bool is_data_type_quantized_per_channel(DataType dt)
+{
+    switch(dt)
+    {
+        case DataType::QSYMM8_PER_CHANNEL:
+            return true;
+        default:
+            return false;
+    }
+}
+
+/** Returns true if the value can be represented by the given data type
+ *
+ * @param[in] val   value to be checked
+ * @param[in] dt    data type that is checked
+ * @param[in] qinfo (Optional) quantization info if the data type is QASYMM8
+ *
+ * @return true if the data type can hold the value.
+ */
+template <typename T>
+bool check_value_range(T val, DataType dt, QuantizationInfo qinfo = QuantizationInfo())
+{
+    switch(dt)
+    {
+        case DataType::U8:
+        {
+            const auto val_u8 = static_cast<uint8_t>(val);
+            return ((val_u8 == val) && val >= std::numeric_limits<uint8_t>::lowest() && val <= std::numeric_limits<uint8_t>::max());
+        }
+        case DataType::QASYMM8:
+        {
+            double min = static_cast<double>(dequantize_qasymm8(0, qinfo));
+            double max = static_cast<double>(dequantize_qasymm8(std::numeric_limits<uint8_t>::max(), qinfo));
+            return ((double)val >= min && (double)val <= max);
+        }
+        case DataType::S8:
+        {
+            const auto val_s8 = static_cast<int8_t>(val);
+            return ((val_s8 == val) && val >= std::numeric_limits<int8_t>::lowest() && val <= std::numeric_limits<int8_t>::max());
+        }
+        case DataType::U16:
+        {
+            const auto val_u16 = static_cast<uint16_t>(val);
+            return ((val_u16 == val) && val >= std::numeric_limits<uint16_t>::lowest() && val <= std::numeric_limits<uint16_t>::max());
+        }
+        case DataType::S16:
+        {
+            const auto val_s16 = static_cast<int16_t>(val);
+            return ((val_s16 == val) && val >= std::numeric_limits<int16_t>::lowest() && val <= std::numeric_limits<int16_t>::max());
+        }
+        case DataType::U32:
+        {
+            const auto val_d64 = static_cast<double>(val);
+            const auto val_u32 = static_cast<uint32_t>(val);
+            return ((val_u32 == val_d64) && val_d64 >= std::numeric_limits<uint32_t>::lowest() && val_d64 <= std::numeric_limits<uint32_t>::max());
+        }
+        case DataType::S32:
+        {
+            const auto val_d64 = static_cast<double>(val);
+            const auto val_s32 = static_cast<int32_t>(val);
+            return ((val_s32 == val_d64) && val_d64 >= std::numeric_limits<int32_t>::lowest() && val_d64 <= std::numeric_limits<int32_t>::max());
+        }
+        case DataType::BFLOAT16:
+            return (val >= bfloat16::lowest() && val <= bfloat16::max());
+        case DataType::F16:
+            return (val >= std::numeric_limits<half>::lowest() && val <= std::numeric_limits<half>::max());
+        case DataType::F32:
+            return (val >= std::numeric_limits<float>::lowest() && val <= std::numeric_limits<float>::max());
+        default:
+            ARM_COMPUTE_ERROR("Data type not supported");
+            return false;
+    }
+}
+
+/** Returns the suffix string of CPU kernel implementation names based on the given data type
+ *
+ * @param[in] data_type The data type the CPU kernel implemetation uses
+ *
+ * @return the suffix string of CPU kernel implementations
+ */
+inline std::string cpu_impl_dt(const DataType &data_type)
+{
+    std::string ret = "";
+
+    switch(data_type)
+    {
+        case DataType::F32:
+            ret = "fp32";
+            break;
+        case DataType::F16:
+            ret = "fp16";
+            break;
+        case DataType::U8:
+            ret = "u8";
+            break;
+        case DataType::S16:
+            ret = "s16";
+            break;
+        case DataType::S32:
+            ret = "s32";
+            break;
+        case DataType::QASYMM8:
+            ret = "qu8";
+            break;
+        case DataType::QASYMM8_SIGNED:
+            ret = "qs8";
+            break;
+        case DataType::QSYMM16:
+            ret = "qs16";
+            break;
+        case DataType::QSYMM8_PER_CHANNEL:
+            ret = "qp8";
+            break;
+        case DataType::BFLOAT16:
+            ret = "bf16";
+            break;
+        default:
+            ARM_COMPUTE_ERROR("Unsupported.");
+    }
+
+    return ret;
+}
+
+}
+#endif /*ARM_COMPUTE_CORE_UTILS_DATATYPEUTILS_H */
diff --git a/arm_compute/core/utils/FormatUtils.h b/arm_compute/core/utils/FormatUtils.h
new file mode 100644
index 0000000000..10e6f747f2
--- /dev/null
+++ b/arm_compute/core/utils/FormatUtils.h
@@ -0,0 +1,344 @@
+/*
+ * Copyright (c) 2016-2023 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_CORE_UTILS_FORMATUTILS_H
+#define ARM_COMPUTE_CORE_UTILS_FORMATUTILS_H
+
+#include "arm_compute/core/Types.h"
+
+namespace arm_compute
+{
+/** The size in bytes of the pixel format
+ *
+ * @param[in] format Input format
+ *
+ * @return The size in bytes of the pixel format
+ */
+inline size_t pixel_size_from_format(Format format)
+{
+    switch(format)
+    {
+        case Format::U8:
+            return 1;
+        case Format::U16:
+        case Format::S16:
+        case Format::BFLOAT16:
+        case Format::F16:
+        case Format::UV88:
+        case Format::YUYV422:
+        case Format::UYVY422:
+            return 2;
+        case Format::RGB888:
+            return 3;
+        case Format::RGBA8888:
+            return 4;
+        case Format::U32:
+        case Format::S32:
+        case Format::F32:
+            return 4;
+        //Doesn't make sense for planar formats:
+        case Format::NV12:
+        case Format::NV21:
+        case Format::IYUV:
+        case Format::YUV444:
+        default:
+            ARM_COMPUTE_ERROR("Undefined pixel size for given format");
+            return 0;
+    }
+}
+
+/** Return the plane index of a given channel given an input format.
+ *
+ * @param[in] format  Input format
+ * @param[in] channel Input channel
+ *
+ * @return The plane index of the specific channel of the specific format
+ */
+inline int plane_idx_from_channel(Format format, Channel channel)
+{
+    switch(format)
+    {
+        // Single planar formats have a single plane
+        case Format::U8:
+        case Format::U16:
+        case Format::S16:
+        case Format::U32:
+        case Format::S32:
+        case Format::BFLOAT16:
+        case Format::F16:
+        case Format::F32:
+        case Format::UV88:
+        case Format::RGB888:
+        case Format::RGBA8888:
+        case Format::YUYV422:
+        case Format::UYVY422:
+            return 0;
+        // Multi planar formats
+        case Format::NV12:
+        case Format::NV21:
+        {
+            // Channel U and V share the same plane of format UV88
+            switch(channel)
+            {
+                case Channel::Y:
+                    return 0;
+                case Channel::U:
+                case Channel::V:
+                    return 1;
+                default:
+                    ARM_COMPUTE_ERROR("Not supported channel");
+                    return 0;
+            }
+        }
+        case Format::IYUV:
+        case Format::YUV444:
+        {
+            switch(channel)
+            {
+                case Channel::Y:
+                    return 0;
+                case Channel::U:
+                    return 1;
+                case Channel::V:
+                    return 2;
+                default:
+                    ARM_COMPUTE_ERROR("Not supported channel");
+                    return 0;
+            }
+        }
+        default:
+            ARM_COMPUTE_ERROR("Not supported format");
+            return 0;
+    }
+}
+
+/** Return the channel index of a given channel given an input format.
+ *
+ * @param[in] format  Input format
+ * @param[in] channel Input channel
+ *
+ * @return The channel index of the specific channel of the specific format
+ */
+inline int channel_idx_from_format(Format format, Channel channel)
+{
+    switch(format)
+    {
+        case Format::RGB888:
+        {
+            switch(channel)
+            {
+                case Channel::R:
+                    return 0;
+                case Channel::G:
+                    return 1;
+                case Channel::B:
+                    return 2;
+                default:
+                    ARM_COMPUTE_ERROR("Not supported channel");
+                    return 0;
+            }
+        }
+        case Format::RGBA8888:
+        {
+            switch(channel)
+            {
+                case Channel::R:
+                    return 0;
+                case Channel::G:
+                    return 1;
+                case Channel::B:
+                    return 2;
+                case Channel::A:
+                    return 3;
+                default:
+                    ARM_COMPUTE_ERROR("Not supported channel");
+                    return 0;
+            }
+        }
+        case Format::YUYV422:
+        {
+            switch(channel)
+            {
+                case Channel::Y:
+                    return 0;
+                case Channel::U:
+                    return 1;
+                case Channel::V:
+                    return 3;
+                default:
+                    ARM_COMPUTE_ERROR("Not supported channel");
+                    return 0;
+            }
+        }
+        case Format::UYVY422:
+        {
+            switch(channel)
+            {
+                case Channel::Y:
+                    return 1;
+                case Channel::U:
+                    return 0;
+                case Channel::V:
+                    return 2;
+                default:
+                    ARM_COMPUTE_ERROR("Not supported channel");
+                    return 0;
+            }
+        }
+        case Format::NV12:
+        {
+            switch(channel)
+            {
+                case Channel::Y:
+                    return 0;
+                case Channel::U:
+                    return 0;
+                case Channel::V:
+                    return 1;
+                default:
+                    ARM_COMPUTE_ERROR("Not supported channel");
+                    return 0;
+            }
+        }
+        case Format::NV21:
+        {
+            switch(channel)
+            {
+                case Channel::Y:
+                    return 0;
+                case Channel::U:
+                    return 1;
+                case Channel::V:
+                    return 0;
+                default:
+                    ARM_COMPUTE_ERROR("Not supported channel");
+                    return 0;
+            }
+        }
+        case Format::YUV444:
+        case Format::IYUV:
+        {
+            switch(channel)
+            {
+                case Channel::Y:
+                    return 0;
+                case Channel::U:
+                    return 0;
+                case Channel::V:
+                    return 0;
+                default:
+                    ARM_COMPUTE_ERROR("Not supported channel");
+                    return 0;
+            }
+        }
+        default:
+            ARM_COMPUTE_ERROR("Not supported format");
+            return 0;
+    }
+}
+
+/** Return the number of planes for a given format
+ *
+ * @param[in] format Input format
+ *
+ * @return The number of planes for a given image format.
+ */
+inline size_t num_planes_from_format(Format format)
+{
+    switch(format)
+    {
+        case Format::U8:
+        case Format::S16:
+        case Format::U16:
+        case Format::S32:
+        case Format::U32:
+        case Format::BFLOAT16:
+        case Format::F16:
+        case Format::F32:
+        case Format::RGB888:
+        case Format::RGBA8888:
+        case Format::YUYV422:
+        case Format::UYVY422:
+            return 1;
+        case Format::NV12:
+        case Format::NV21:
+            return 2;
+        case Format::IYUV:
+        case Format::YUV444:
+            return 3;
+        default:
+            ARM_COMPUTE_ERROR("Not supported format");
+            return 0;
+    }
+}
+
+/** Return the number of channels for a given single-planar pixel format
+ *
+ * @param[in] format Input format
+ *
+ * @return The number of channels for a given image format.
+ */
+inline size_t num_channels_from_format(Format format)
+{
+    switch(format)
+    {
+        case Format::U8:
+        case Format::U16:
+        case Format::S16:
+        case Format::U32:
+        case Format::S32:
+        case Format::BFLOAT16:
+        case Format::F16:
+        case Format::F32:
+            return 1;
+        // Because the U and V channels are subsampled
+        // these formats appear like having only 2 channels:
+        case Format::YUYV422:
+        case Format::UYVY422:
+            return 2;
+        case Format::UV88:
+            return 2;
+        case Format::RGB888:
+            return 3;
+        case Format::RGBA8888:
+            return 4;
+        //Doesn't make sense for planar formats:
+        case Format::NV12:
+        case Format::NV21:
+        case Format::IYUV:
+        case Format::YUV444:
+        default:
+            return 0;
+    }
+}
+
+/** Convert a tensor format into a string.
+ *
+ * @param[in] format @ref Format to be translated to string.
+ *
+ * @return The string describing the format.
+ */
+const std::string &string_from_format(Format format);
+
+}
+#endif /*ARM_COMPUTE_CORE_UTILS_FORMATUTILS_H */
diff --git a/arm_compute/core/utils/InterpolationPolicyUtils.h b/arm_compute/core/utils/InterpolationPolicyUtils.h
new file mode 100644
index 0000000000..79f6e3aa5f
--- /dev/null
+++ b/arm_compute/core/utils/InterpolationPolicyUtils.h
@@ -0,0 +1,41 @@
+/*
+ * Copyright (c) 2016-2023 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_CORE_UTILS_INTERPOLATIONPOLICYUTILS_H
+#define ARM_COMPUTE_CORE_UTILS_INTERPOLATIONPOLICYUTILS_H
+
+#include "arm_compute/core/Types.h"
+
+#include <string>
+
+namespace arm_compute
+{
+/** Translates a given interpolation policy to a string.
+ *
+ * @param[in] policy @ref InterpolationPolicy to be translated to string.
+ *
+ * @return The string describing the interpolation policy.
+ */
+const std::string &string_from_interpolation_policy(InterpolationPolicy policy);
+}
+#endif /*ARM_COMPUTE_CORE_UTILS_INTERPOLATIONPOLICYUTILS_H */
diff --git a/arm_compute/core/utils/StringUtils.h b/arm_compute/core/utils/StringUtils.h
new file mode 100644
index 0000000000..a91ef2387d
--- /dev/null
+++ b/arm_compute/core/utils/StringUtils.h
@@ -0,0 +1,55 @@
+/*
+ * Copyright (c) 2016-2023 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_CORE_UTILS_STRINGUTILS_H
+#define ARM_COMPUTE_CORE_UTILS_STRINGUTILS_H
+
+#include <string>
+
+namespace arm_compute
+{
+/** Lower a given string.
+ *
+ * @param[in] val Given string to lower.
+ *
+ * @return The lowered string
+ */
+std::string lower_string(const std::string &val);
+
+/** Raise a given string to upper case
+ *
+ * @param[in] val Given string to lower.
+ *
+ * @return The upper case string
+ */
+std::string upper_string(const std::string &val);
+
+/** Create a string with the float in full precision.
+ *
+ * @param val Floating point value
+ *
+ * @return String with the floating point value.
+ */
+std::string float_to_string_with_full_precision(float val);
+}
+#endif /*ARM_COMPUTE_CORE_UTILS_STRINGUTILS_H */
diff --git a/arm_compute/core/utils/helpers/AdjustVecSize.h b/arm_compute/core/utils/helpers/AdjustVecSize.h
new file mode 100644
index 0000000000..bbb3048b84
--- /dev/null
+++ b/arm_compute/core/utils/helpers/AdjustVecSize.h
@@ -0,0 +1,55 @@
+/*
+ * Copyright (c) 2016-2023 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_UTILS_ADJUSTVECSIZE_H
+#define ARM_COMPUTE_UTILS_ADJUSTVECSIZE_H
+
+#include "arm_compute/core/Error.h"
+
+namespace arm_compute
+{
+/** Returns the adjusted vector size in case it is less than the input's first dimension, getting rounded down to its closest valid vector size
+ *
+ * @param[in] vec_size vector size to be adjusted
+ * @param[in] dim0     size of the first dimension
+ *
+ * @return the number of element processed along the X axis per thread
+ */
+inline unsigned int adjust_vec_size(unsigned int vec_size, size_t dim0)
+{
+    ARM_COMPUTE_ERROR_ON(vec_size > 16);
+
+    if((vec_size >= dim0) && (dim0 == 3))
+    {
+        return dim0;
+    }
+
+    while(vec_size > dim0)
+    {
+        vec_size >>= 1;
+    }
+
+    return vec_size;
+}
+}
+#endif /*ARM_COMPUTE_UTILS_H */