1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
|
/*
* Copyright (c) 2018 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_TEST_HOG_MULTI_DETECTION_FIXTURE
#define ARM_COMPUTE_TEST_HOG_MULTI_DETECTION_FIXTURE
#include "arm_compute/core/HOGInfo.h"
#include "arm_compute/core/TensorShape.h"
#include "arm_compute/core/Types.h"
#include "tests/AssetsLibrary.h"
#include "tests/Globals.h"
#include "tests/IAccessor.h"
#include "tests/IHOGAccessor.h"
#include "tests/framework/Asserts.h"
#include "tests/framework/Fixture.h"
#include "tests/validation/reference/HOGMultiDetection.h"
namespace arm_compute
{
namespace test
{
namespace validation
{
template <typename TensorType,
typename HOGType,
typename MultiHOGType,
typename DetectionWindowArrayType,
typename DetectionWindowStrideType,
typename AccessorType,
typename Size2DArrayAccessorType,
typename DetectionWindowArrayAccessorType,
typename HOGAccessorType,
typename FunctionType,
typename T,
typename U>
class HOGMultiDetectionValidationFixture : public framework::Fixture
{
public:
template <typename...>
void setup(std::string image, std::vector<HOGInfo> models, Format format, BorderMode border_mode, bool non_maxima_suppression)
{
// Only defined borders supported
ARM_COMPUTE_ERROR_ON(border_mode == BorderMode::UNDEFINED);
// Generate a random constant value
std::mt19937 gen(library->seed());
std::uniform_int_distribution<T> int_dist(0, 255);
const T constant_border_value = int_dist(gen);
// Initialize descriptors vector
std::vector<std::vector<U>> descriptors(models.size());
// Use default values for threshold and min_distance
const float threshold = 0.f;
const float min_distance = 1.f;
// Maximum number of detection windows per batch
const unsigned int max_num_detection_windows = 100000;
_target = compute_target(image, format, border_mode, constant_border_value, models, descriptors, max_num_detection_windows, threshold, non_maxima_suppression, min_distance);
_reference = compute_reference(image, format, border_mode, constant_border_value, models, descriptors, max_num_detection_windows, threshold, non_maxima_suppression, min_distance);
}
protected:
template <typename V>
void fill(V &&tensor, const std::string image, Format format)
{
library->fill(tensor, image, format);
}
void initialize_batch(const std::vector<HOGInfo> &models, MultiHOGType &multi_hog,
std::vector<std::vector<U>> &descriptors, DetectionWindowStrideType &detection_window_strides)
{
for(unsigned i = 0; i < models.size(); ++i)
{
auto hog_model = reinterpret_cast<HOGType *>(multi_hog.model(i));
hog_model->init(models[i]);
// Initialise descriptor (linear SVM coefficients).
std::random_device::result_type seed = 0;
descriptors.at(i) = generate_random_real(models[i].descriptor_size(), -0.505f, 0.495f, seed);
// Copy HOG descriptor values to HOG memory
{
HOGAccessorType hog_accessor(*hog_model);
std::memcpy(hog_accessor.descriptor(), descriptors.at(i).data(), descriptors.at(i).size() * sizeof(U));
}
// Initialize detection window stride
Size2DArrayAccessorType accessor(detection_window_strides);
accessor.at(i) = models[i].block_stride();
}
}
std::vector<DetectionWindow> compute_target(const std::string image, Format &format, BorderMode &border_mode, T constant_border_value,
const std::vector<HOGInfo> &models, std::vector<std::vector<U>> &descriptors, unsigned int max_num_detection_windows,
float threshold, bool non_max_suppression, float min_distance)
{
MultiHOGType multi_hog(models.size());
DetectionWindowArrayType detection_windows(max_num_detection_windows);
DetectionWindowStrideType detection_window_strides(models.size());
// Resize detection window_strides for index access
detection_window_strides.resize(models.size());
// Initialiize MultiHOG and detection windows
initialize_batch(models, multi_hog, descriptors, detection_window_strides);
// Get image shape for src tensor
TensorShape shape = library->get_image_shape(image);
// Create tensors
TensorType src = create_tensor<TensorType>(shape, data_type_from_format(format));
ARM_COMPUTE_EXPECT(src.info()->is_resizable(), framework::LogLevel::ERRORS);
// Create and configure function
FunctionType hog_multi_detection;
hog_multi_detection.configure(&src, &multi_hog, &detection_windows, &detection_window_strides, border_mode, constant_border_value, threshold, non_max_suppression, min_distance);
// Reset detection windows
detection_windows.clear();
// Allocate tensors
src.allocator()->allocate();
ARM_COMPUTE_EXPECT(!src.info()->is_resizable(), framework::LogLevel::ERRORS);
// Fill tensors
fill(AccessorType(src), image, format);
// Compute function
hog_multi_detection.run();
// Copy detection windows
std::vector<DetectionWindow> windows;
DetectionWindowArrayAccessorType accessor(detection_windows);
for(size_t i = 0; i < accessor.num_values(); i++)
{
DetectionWindow win;
win.x = accessor.at(i).x;
win.y = accessor.at(i).y;
win.width = accessor.at(i).width;
win.height = accessor.at(i).height;
win.idx_class = accessor.at(i).idx_class;
win.score = accessor.at(i).score;
windows.push_back(win);
}
return windows;
}
std::vector<DetectionWindow> compute_reference(const std::string image, Format format, BorderMode border_mode, T constant_border_value,
const std::vector<HOGInfo> &models, const std::vector<std::vector<U>> &descriptors, unsigned int max_num_detection_windows,
float threshold, bool non_max_suppression, float min_distance)
{
// Create reference
SimpleTensor<T> src{ library->get_image_shape(image), data_type_from_format(format) };
// Fill reference
fill(src, image, format);
// NOTE: Detection window stride fixed to block stride
return reference::hog_multi_detection(src, border_mode, constant_border_value, models, descriptors, max_num_detection_windows, threshold, non_max_suppression, min_distance);
}
std::vector<DetectionWindow> _target{};
std::vector<DetectionWindow> _reference{};
};
} // namespace validation
} // namespace test
} // namespace arm_compute
#endif /* ARM_COMPUTE_TEST_HOG_MULTI_DETECTION_FIXTURE */
|
