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
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
/*M///////////////////////////////////////////////////////////////////////////////////////
//
// IMPORTANT: READ BEFORE DOWNLOADING, COPYING, INSTALLING OR USING.
//
// By downloading, copying, installing or using the software you agree to this license.
// If you do not agree to this license, do not download, install,
// copy or use the software.
//
//
// License Agreement
// For Open Source Computer Vision Library
//
// Copyright (C) 2013, OpenCV Foundation, all rights reserved.
// Copyright (C) 2017, Intel Corporation, all rights reserved.
// Third party copyrights are property of their respective owners.
//
// Redistribution and use in source and binary forms, with or without modification,
// are permitted provided that the following conditions are met:
//
// * Redistribution's of source code must retain the above copyright notice,
// this list of conditions and the following disclaimer.
//
// * Redistribution's in binary form must reproduce the above copyright notice,
// this list of conditions and the following disclaimer in the documentation
// and/or other materials provided with the distribution.
//
// * The name of the copyright holders may not be used to endorse or promote products
// derived from this software without specific prior written permission.
//
// This software is provided by the copyright holders and contributors "as is" and
// any express or implied warranties, including, but not limited to, the implied
// warranties of merchantability and fitness for a particular purpose are disclaimed.
// In no event shall the Intel Corporation or contributors be liable for any direct,
// indirect, incidental, special, exemplary, or consequential damages
// (including, but not limited to, procurement of substitute goods or services;
// loss of use, data, or profits; or business interruption) however caused
// and on any theory of liability, whether in contract, strict liability,
// or tort (including negligence or otherwise) arising in any way out of
// the use of this software, even if advised of the possibility of such damage.
//
//M*/
#include "../precomp.hpp"
#include "layers_common.hpp"
namespace cv
{
namespace dnn
{
namespace util
{
std::string makeName(const std::string& str1, const std::string& str2)
{
return str1 + str2;
}
bool getParameter(const LayerParams ¶ms, const std::string& nameBase, const std::string& nameAll,
std::vector<size_t>& parameter, bool hasDefault = false, const std::vector<size_t>& defaultValue = std::vector<size_t>(2, 0))
{
std::string nameH = makeName(nameBase, std::string("_h"));
std::string nameW = makeName(nameBase, std::string("_w"));
std::string nameAll_ = nameAll;
if (nameAll_ == "")
nameAll_ = nameBase;
if (params.has(nameH) && params.has(nameW))
{
CV_Assert(params.get<int>(nameH) >= 0 && params.get<int>(nameW) >= 0);
parameter.push_back(params.get<int>(nameH));
parameter.push_back(params.get<int>(nameW));
return true;
}
else
{
if (params.has(nameAll_))
{
DictValue param = params.get(nameAll_);
for (int i = 0; i < param.size(); i++) {
CV_Assert(param.get<int>(i) >= 0);
parameter.push_back(param.get<int>(i));
}
if (parameter.size() == 1)
parameter.resize(2, parameter[0]);
return true;
}
else
{
if (hasDefault)
{
parameter = defaultValue;
return true;
}
else
{
return false;
}
}
}
}
void getKernelSize(const LayerParams ¶ms, std::vector<size_t>& kernel)
{
if (!util::getParameter(params, "kernel", "kernel_size", kernel))
CV_Error(cv::Error::StsBadArg, "kernel_size (or kernel_h and kernel_w) not specified");
for (int i = 0; i < kernel.size(); i++)
CV_Assert(kernel[i] > 0);
}
void getStrideAndPadding(const LayerParams ¶ms, std::vector<size_t>& pads_begin, std::vector<size_t>& pads_end,
std::vector<size_t>& strides, cv::String& padMode, size_t kernel_size = 2)
{
if (params.has("pad_l") && params.has("pad_t") && params.has("pad_r") && params.has("pad_b")) {
CV_Assert(params.get<int>("pad_t") >= 0 && params.get<int>("pad_l") >= 0 &&
params.get<int>("pad_b") >= 0 && params.get<int>("pad_r") >= 0);
pads_begin.push_back(params.get<int>("pad_t"));
pads_begin.push_back(params.get<int>("pad_l"));
pads_end.push_back(params.get<int>("pad_b"));
pads_end.push_back(params.get<int>("pad_r"));
}
else {
util::getParameter(params, "pad", "pad", pads_begin, true, std::vector<size_t>(kernel_size, 0));
if (pads_begin.size() < 4)
pads_end = pads_begin;
else
{
pads_end = std::vector<size_t>(pads_begin.begin() + pads_begin.size() / 2, pads_begin.end());
pads_begin.resize(pads_begin.size() / 2);
}
CV_Assert(pads_begin.size() == pads_end.size());
}
util::getParameter(params, "stride", "stride", strides, true, std::vector<size_t>(kernel_size, 1));
padMode = "";
if (params.has("pad_mode"))
{
padMode = params.get<String>("pad_mode");
}
for (int i = 0; i < strides.size(); i++)
CV_Assert(strides[i] > 0);
}
}
void getPoolingKernelParams(const LayerParams ¶ms, std::vector<size_t>& kernel, bool &globalPooling,
std::vector<size_t>& pads_begin, std::vector<size_t>& pads_end,
std::vector<size_t>& strides, cv::String &padMode)
{
globalPooling = params.has("global_pooling") &&
params.get<bool>("global_pooling");
if (globalPooling)
{
util::getStrideAndPadding(params, pads_begin, pads_end, strides, padMode);
if(params.has("kernel_h") || params.has("kernel_w") || params.has("kernel_size"))
{
CV_Error(cv::Error::StsBadArg, "In global_pooling mode, kernel_size (or kernel_h and kernel_w) cannot be specified");
}
for (int i = 0; i < pads_begin.size(); i++) {
if (pads_begin[i] != 0 || pads_end[i] != 0)
CV_Error(cv::Error::StsBadArg, "In global_pooling mode, pads must be = 0");
}
for (int i = 0; i < strides.size(); i++) {
if (strides[i] != 1)
CV_Error(cv::Error::StsBadArg, "In global_pooling mode, strides must be = 1");
}
}
else
{
util::getKernelSize(params, kernel);
util::getStrideAndPadding(params, pads_begin, pads_end, strides, padMode, kernel.size());
}
}
void getConvolutionKernelParams(const LayerParams ¶ms, std::vector<size_t>& kernel, std::vector<size_t>& pads_begin,
std::vector<size_t>& pads_end, std::vector<size_t>& strides,
std::vector<size_t>& dilations, cv::String &padMode, std::vector<size_t>& adjust_pads)
{
util::getKernelSize(params, kernel);
util::getStrideAndPadding(params, pads_begin, pads_end, strides, padMode, kernel.size());
util::getParameter(params, "dilation", "dilation", dilations, true, std::vector<size_t>(kernel.size(), 1));
util::getParameter(params, "adj", "adj", adjust_pads, true, std::vector<size_t>(kernel.size(), 0));
for (int i = 0; i < dilations.size(); i++)
CV_Assert(dilations[i] > 0);
}
// From TensorFlow code:
// Total padding on rows and cols is
// Pr = (R' - 1) * S + Kr - R
// Pc = (C' - 1) * S + Kc - C
// where (R', C') are output dimensions, (R, C) are input dimensions, S
// is stride, (Kr, Kc) are filter dimensions.
// We pad Pr/2 on the left and Pr - Pr/2 on the right, Pc/2 on the top
// and Pc - Pc/2 on the bottom. When Pr or Pc is odd, this means
// we pad more on the right and bottom than on the top and left.
void getConvPoolOutParams(const std::vector<int>& inp, const std::vector<size_t>& kernel,
const std::vector<size_t>& stride, const String &padMode,
const std::vector<size_t>& dilation, std::vector<int>& out)
{
if (padMode == "VALID")
{
for (int i = 0; i < inp.size(); i++)
out.push_back((inp[i] - dilation[i] * (kernel[i] - 1) - 1 + stride[i]) / stride[i]);
}
else if (padMode == "SAME")
{
for (int i = 0; i < inp.size(); i++)
out.push_back((inp[i] - 1 + stride[i]) / stride[i]);
}
else
{
CV_Error(Error::StsError, "Unsupported padding mode");
}
}
void getConvPoolPaddings(const std::vector<int>& inp, const std::vector<size_t>& kernel,
const std::vector<size_t>& strides, const String &padMode,
std::vector<size_t>& pads_begin, std::vector<size_t>& pads_end)
{
if (padMode == "SAME" || padMode == "VALID")
{
pads_begin.assign(kernel.size(), 0);
pads_end.assign(kernel.size(), 0);
}
if (padMode == "SAME")
{
CV_Assert_N(kernel.size() == strides.size(), kernel.size() == inp.size());
for (int i = 0; i < pads_begin.size(); i++) {
// There are test cases with stride > kernel.
if (strides[i] <= kernel[i])
{
int pad = (kernel[i] - 1 - (inp[i] - 1 + strides[i]) % strides[i]) / 2;
pads_begin[i] = pads_end[i] = pad;
}
}
}
}
}
}