potok/opencv
1
0
Fork 0
mirror of https://github.com/opencv/opencv.git synced 2026-07-31 08:13:04 +04:00
Code Packages Releases Wiki Activity

Merge remote-tracking branch 'upstream/3.4' into merge-3.4

Browse Source
This commit is contained in:
Alexander Alekhin
2019-05-28 19:41:48 +03:00
committed by Alexander Alekhin
parent ed96864892 0d477d7364
commit 43467a2ac7
56 changed files with 2025 additions and 1195 deletions
Download Patch File Download Diff File Expand all files Collapse all files
modules/dnn/include/opencv2/dnn/all_layers.hpp
+1 -1
View File
@@ -608,7 +608,7 @@ CV__DNN_INLINE_NS_BEGIN
};
/**
* @brief Bilinear resize layer from https://github.com/cdmh/deeplab-public
* @brief Bilinear resize layer from https://github.com/cdmh/deeplab-public-ver2
*
* It differs from @ref ResizeLayer in output shape and resize scales computations.
*/
modules/dnn/perf/perf_net.cpp
+1 -2
View File
@@ -214,8 +214,7 @@ PERF_TEST_P_(DNNTestNetwork, EAST_text_detection)
PERF_TEST_P_(DNNTestNetwork, FastNeuralStyle_eccv16)
{
if (backend == DNN_BACKEND_HALIDE ||
(backend == DNN_BACKEND_OPENCV && target == DNN_TARGET_OPENCL_FP16) ||
(backend == DNN_BACKEND_INFERENCE_ENGINE && target == DNN_TARGET_MYRIAD))
(backend == DNN_BACKEND_OPENCV && target == DNN_TARGET_OPENCL_FP16))
throw SkipTestException("");
processNet("dnn/fast_neural_style_eccv16_starry_night.t7", "", "", Mat(cv::Size(320, 240), CV_32FC3));
}
modules/dnn/src/dnn.cpp
+53 -28
View File
@@ -3056,6 +3056,23 @@ int Net::getLayerId(const String &layer)
return impl->getLayerId(layer);
}
String parseLayerParams(const String& name, const LayerParams& lp) {
DictValue param = lp.get(name);
std::ostringstream out;
out << name << " ";
switch (param.size()) {
case 1: out << ": "; break;
case 2: out << "(HxW): "; break;
case 3: out << "(DxHxW): "; break;
default: CV_Error(Error::StsNotImplemented, format("Unsupported %s size = %d", name.c_str(), param.size()));
}
for (size_t i = 0; i < param.size() - 1; i++) {
out << param.get<int>(i) << " x ";
}
out << param.get<int>(param.size() - 1) << "\\l";
return out.str();
}
String Net::dump()
{
CV_Assert(!empty());
@@ -3141,39 +3158,47 @@ String Net::dump()
out << " | ";
}
out << lp.name << "\\n" << lp.type << "\\n";
if (lp.has("kernel_size")) {
DictValue size = lp.get("kernel_size");
out << "kernel (HxW): " << size << " x " << size << "\\l";
} else if (lp.has("kernel_h") && lp.has("kernel_w")) {
DictValue h = lp.get("kernel_h");
DictValue w = lp.get("kernel_w");
out << "kernel (HxW): " << h << " x " << w << "\\l";
}
if (lp.has("stride")) {
DictValue stride = lp.get("stride");
out << "stride (HxW): " << stride << " x " << stride << "\\l";
} else if (lp.has("stride_h") && lp.has("stride_w")) {
DictValue h = lp.get("stride_h");
DictValue w = lp.get("stride_w");
out << "stride (HxW): " << h << " x " << w << "\\l";
}
if (lp.has("dilation")) {
DictValue dilation = lp.get("dilation");
out << "dilation (HxW): " << dilation << " x " << dilation << "\\l";
} else if (lp.has("dilation_h") && lp.has("dilation_w")) {
DictValue h = lp.get("dilation_h");
DictValue w = lp.get("dilation_w");
out << "dilation (HxW): " << h << " x " << w << "\\l";
}
if (lp.has("pad")) {
DictValue pad = lp.get("pad");
out << "pad (LxTxRxB): " << pad << " x " << pad << " x " << pad << " x " << pad << "\\l";
if (lp.has("kernel_size")) {
String kernel = parseLayerParams("kernel_size", lp);
out << kernel;
} else if (lp.has("kernel_h") && lp.has("kernel_w")) {
DictValue h = lp.get("kernel_h");
DictValue w = lp.get("kernel_w");
out << "kernel (HxW): " << h << " x " << w << "\\l";
}
if (lp.has("stride")) {
String stride = parseLayerParams("stride", lp);
out << stride;
} else if (lp.has("stride_h") && lp.has("stride_w")) {
DictValue h = lp.get("stride_h");
DictValue w = lp.get("stride_w");
out << "stride (HxW): " << h << " x " << w << "\\l";
}
if (lp.has("dilation")) {
String dilation = parseLayerParams("dilation", lp);
out << dilation;
} else if (lp.has("dilation_h") && lp.has("dilation_w")) {
DictValue h = lp.get("dilation_h");
DictValue w = lp.get("dilation_w");
out << "dilation (HxW): " << h << " x " << w << "\\l";
}
if (lp.has("pad")) {
DictValue pad = lp.get("pad");
out << "pad ";
switch (pad.size()) {
case 1: out << ": " << pad << "\\l"; break;
case 2: out << "(HxW): (" << pad.get<int>(0) << " x " << pad.get<int>(1) << ")" << "\\l"; break;
case 4: out << "(HxW): (" << pad.get<int>(0) << ", " << pad.get<int>(2) << ") x (" << pad.get<int>(1) << ", " << pad.get<int>(3) << ")" << "\\l"; break;
case 6: out << "(DxHxW): (" << pad.get<int>(0) << ", " << pad.get<int>(3) << ") x (" << pad.get<int>(1) << ", " << pad.get<int>(4)
<< ") x (" << pad.get<int>(2) << ", " << pad.get<int>(5) << ")" << "\\l"; break;
default: CV_Error(Error::StsNotImplemented, format("Unsupported pad size = %d", pad.size()));
}
} else if (lp.has("pad_l") && lp.has("pad_t") && lp.has("pad_r") && lp.has("pad_b")) {
DictValue l = lp.get("pad_l");
DictValue t = lp.get("pad_t");
DictValue r = lp.get("pad_r");
DictValue b = lp.get("pad_b");
out << "pad (LxTxRxB): " << l << " x " << t << " x " << r << " x " << b << "\\l";
out << "pad (HxW): (" << t << ", " << b << ") x (" << l << ", " << r << ")" << "\\l";
}
else if (lp.has("pooled_w") || lp.has("pooled_h")) {
DictValue h = lp.get("pooled_h");
modules/dnn/src/layers/slice_layer.cpp
+43 -23
View File
@@ -110,15 +110,9 @@ public:
virtual bool supportBackend(int backendId) CV_OVERRIDE
{
#ifdef HAVE_INF_ENGINE
if (backendId == DNN_BACKEND_INFERENCE_ENGINE)
{
return INF_ENGINE_VER_MAJOR_LT(INF_ENGINE_RELEASE_2018R5) &&
sliceRanges.size() == 1 && sliceRanges[0].size() == 4;
}
else
#endif
return backendId == DNN_BACKEND_OPENCV;
return backendId == DNN_BACKEND_OPENCV ||
(backendId == DNN_BACKEND_INFERENCE_ENGINE &&
sliceRanges.size() == 1 && sliceRanges[0].size() == 4);
}
bool getMemoryShapes(const std::vector<MatShape> &inputs,
@@ -264,39 +258,65 @@ public:
#ifdef HAVE_INF_ENGINE
virtual Ptr<BackendNode> initInfEngine(const std::vector<Ptr<BackendWrapper> >& inputs) CV_OVERRIDE
{
#if INF_ENGINE_VER_MAJOR_LT(INF_ENGINE_RELEASE_2018R5)
InferenceEngine::DataPtr input = infEngineDataNode(inputs[0]);
InferenceEngine::LayerParams lp;
lp.name = name;
lp.type = "Crop";
lp.precision = InferenceEngine::Precision::FP32;
std::shared_ptr<InferenceEngine::CropLayer> ieLayer(new InferenceEngine::CropLayer(lp));
CV_Assert(sliceRanges.size() == 1);
std::vector<size_t> axes, offsets, dims;
int from, to, step;
int numDims = sliceRanges[0].size();
if (preferableTarget == DNN_TARGET_MYRIAD)
{
from = 1;
to = sliceRanges[0].size() + 1;
to = numDims;
step = 1;
}
else
{
from = sliceRanges[0].size() - 1;
from = numDims - 1;
to = -1;
step = -1;
}
for (int i = from; i != to; i += step)
{
ieLayer->axis.push_back(i);
ieLayer->offset.push_back(sliceRanges[0][i].start);
ieLayer->dim.push_back(sliceRanges[0][i].end - sliceRanges[0][i].start);
axes.push_back(i);
offsets.push_back(sliceRanges[0][i].start);
dims.push_back(sliceRanges[0][i].size());
}
#if INF_ENGINE_VER_MAJOR_GE(INF_ENGINE_RELEASE_2018R5)
std::vector<size_t> outShape(numDims);
for (int i = 0; i < numDims; ++i)
outShape[numDims - 1 - i] = sliceRanges[0][i].size();
InferenceEngine::Builder::Layer ieLayer(name);
ieLayer.setName(name);
ieLayer.setType("Crop");
ieLayer.getParameters()["axis"] = axes;
ieLayer.getParameters()["dim"] = dims;
ieLayer.getParameters()["offset"] = offsets;
ieLayer.setInputPorts(std::vector<InferenceEngine::Port>(2));
ieLayer.setOutputPorts(std::vector<InferenceEngine::Port>(1));
ieLayer.getInputPorts()[1].setParameter("type", "weights");
// Fake blob which will be moved to inputs (as weights).
auto shapeSource = InferenceEngine::make_shared_blob<float>(
InferenceEngine::Precision::FP32,
InferenceEngine::Layout::ANY, outShape);
shapeSource->allocate();
addConstantData("weights", shapeSource, ieLayer);
return Ptr<BackendNode>(new InfEngineBackendNode(ieLayer));
#else
return Ptr<BackendNode>();
InferenceEngine::LayerParams lp;
lp.name = name;
lp.type = "Crop";
lp.precision = InferenceEngine::Precision::FP32;
std::shared_ptr<InferenceEngine::CropLayer> ieLayer(new InferenceEngine::CropLayer(lp));
ieLayer->axis = axes;
ieLayer->offset = offsets;
ieLayer->dim = dims;
return Ptr<BackendNode>(new InfEngineBackendNode(ieLayer));
#endif // IE < R5
return Ptr<BackendNode>();
}
#endif
};
modules/dnn/test/test_backends.cpp
+11 -3
View File
@@ -110,7 +110,10 @@ TEST_P(DNNTestNetwork, AlexNet)
TEST_P(DNNTestNetwork, ResNet_50)
{
applyTestTag(target == DNN_TARGET_CPU ? CV_TEST_TAG_MEMORY_512MB : CV_TEST_TAG_MEMORY_1GB);
applyTestTag(
(target == DNN_TARGET_CPU ? CV_TEST_TAG_MEMORY_512MB : CV_TEST_TAG_MEMORY_1GB),
CV_TEST_TAG_DEBUG_LONG
);
processNet("dnn/ResNet-50-model.caffemodel", "dnn/ResNet-50-deploy.prototxt",
Size(224, 224), "prob",
target == DNN_TARGET_OPENCL ? "dnn/halide_scheduler_opencl_resnet_50.yml" :
@@ -344,7 +347,10 @@ TEST_P(DNNTestNetwork, opencv_face_detector)
TEST_P(DNNTestNetwork, Inception_v2_SSD_TensorFlow)
{
applyTestTag(target == DNN_TARGET_CPU ? CV_TEST_TAG_MEMORY_512MB : CV_TEST_TAG_MEMORY_1GB);
applyTestTag(
(target == DNN_TARGET_CPU ? CV_TEST_TAG_MEMORY_512MB : CV_TEST_TAG_MEMORY_1GB),
CV_TEST_TAG_DEBUG_LONG
);
#if defined(INF_ENGINE_RELEASE)
if (backend == DNN_BACKEND_INFERENCE_ENGINE && target == DNN_TARGET_MYRIAD
&& getInferenceEngineVPUType() == CV_DNN_INFERENCE_ENGINE_VPU_TYPE_MYRIAD_X)
@@ -382,6 +388,8 @@ TEST_P(DNNTestNetwork, DenseNet_121)
TEST_P(DNNTestNetwork, FastNeuralStyle_eccv16)
{
applyTestTag(CV_TEST_TAG_MEMORY_512MB, CV_TEST_TAG_DEBUG_VERYLONG);
if (backend == DNN_BACKEND_HALIDE ||
(backend == DNN_BACKEND_INFERENCE_ENGINE && target == DNN_TARGET_MYRIAD))
throw SkipTestException("");
@@ -396,7 +404,7 @@ TEST_P(DNNTestNetwork, FastNeuralStyle_eccv16)
Mat img = imread(findDataFile("dnn/googlenet_1.png", false));
Mat inp = blobFromImage(img, 1.0, Size(320, 240), Scalar(103.939, 116.779, 123.68), false, false);
// Output image has values in range [-143.526, 148.539].
float l1 = (target == DNN_TARGET_OPENCL_FP16 || target == DNN_TARGET_MYRIAD) ? 0.3 : 4e-5;
float l1 = (target == DNN_TARGET_OPENCL_FP16 || target == DNN_TARGET_MYRIAD) ? 0.4 : 4e-5;
float lInf = (target == DNN_TARGET_OPENCL_FP16 || target == DNN_TARGET_MYRIAD) ? 7.28 : 2e-3;
processNet("dnn/fast_neural_style_eccv16_starry_night.t7", "", inp, "", "", l1, lInf);
}
modules/dnn/test/test_caffe_importer.cpp
+23 -5
View File
@@ -114,6 +114,9 @@ TEST_P(Reproducibility_AlexNet, Accuracy)
{
Target targetId = get<1>(GetParam());
applyTestTag(targetId == DNN_TARGET_CPU ? CV_TEST_TAG_MEMORY_512MB : CV_TEST_TAG_MEMORY_1GB);
if (!ocl::useOpenCL() && targetId != DNN_TARGET_CPU)
throw SkipTestException("OpenCL is disabled");
bool readFromMemory = get<0>(GetParam());
Net net;
{
@@ -154,7 +157,8 @@ INSTANTIATE_TEST_CASE_P(/**/, Reproducibility_AlexNet, Combine(testing::Bool(),
TEST(Reproducibility_FCN, Accuracy)
{
applyTestTag(CV_TEST_TAG_LONG, CV_TEST_TAG_MEMORY_2GB);
applyTestTag(CV_TEST_TAG_LONG, CV_TEST_TAG_DEBUG_VERYLONG, CV_TEST_TAG_MEMORY_2GB);
Net net;
{
const string proto = findDataFile("dnn/fcn8s-heavy-pascal.prototxt", false);
@@ -183,7 +187,7 @@ TEST(Reproducibility_FCN, Accuracy)
TEST(Reproducibility_SSD, Accuracy)
{
applyTestTag(CV_TEST_TAG_MEMORY_512MB);
applyTestTag(CV_TEST_TAG_MEMORY_512MB, CV_TEST_TAG_DEBUG_LONG);
Net net;
{
const string proto = findDataFile("dnn/ssd_vgg16.prototxt", false);
@@ -281,6 +285,9 @@ TEST_P(Reproducibility_ResNet50, Accuracy)
{
Target targetId = GetParam();
applyTestTag(targetId == DNN_TARGET_CPU ? CV_TEST_TAG_MEMORY_512MB : CV_TEST_TAG_MEMORY_1GB);
if (!ocl::useOpenCL() && targetId != DNN_TARGET_CPU)
throw SkipTestException("OpenCL is disabled");
Net net = readNetFromCaffe(findDataFile("dnn/ResNet-50-deploy.prototxt", false),
findDataFile("dnn/ResNet-50-model.caffemodel", false));
@@ -541,7 +548,11 @@ INSTANTIATE_TEST_CASE_P(Test_Caffe, opencv_face_detector,
TEST_P(Test_Caffe_nets, FasterRCNN_vgg16)
{
applyTestTag(CV_TEST_TAG_LONG, (target == DNN_TARGET_CPU ? CV_TEST_TAG_MEMORY_1GB : CV_TEST_TAG_MEMORY_2GB));
applyTestTag(
(target == DNN_TARGET_CPU ? CV_TEST_TAG_MEMORY_1GB : CV_TEST_TAG_MEMORY_2GB),
CV_TEST_TAG_LONG,
CV_TEST_TAG_DEBUG_VERYLONG
);
#if defined(INF_ENGINE_RELEASE)
if (backend == DNN_BACKEND_INFERENCE_ENGINE && (target == DNN_TARGET_OPENCL || target == DNN_TARGET_OPENCL_FP16))
@@ -559,7 +570,10 @@ TEST_P(Test_Caffe_nets, FasterRCNN_vgg16)
TEST_P(Test_Caffe_nets, FasterRCNN_zf)
{
applyTestTag(target == DNN_TARGET_CPU ? CV_TEST_TAG_MEMORY_512MB : CV_TEST_TAG_MEMORY_1GB);
applyTestTag(
(target == DNN_TARGET_CPU ? CV_TEST_TAG_MEMORY_512MB : CV_TEST_TAG_MEMORY_1GB),
CV_TEST_TAG_DEBUG_LONG
);
if ((backend == DNN_BACKEND_INFERENCE_ENGINE && target == DNN_TARGET_OPENCL_FP16) ||
(backend == DNN_BACKEND_INFERENCE_ENGINE && target == DNN_TARGET_MYRIAD))
throw SkipTestException("");
@@ -571,7 +585,11 @@ TEST_P(Test_Caffe_nets, FasterRCNN_zf)
TEST_P(Test_Caffe_nets, RFCN)
{
applyTestTag(CV_TEST_TAG_LONG, (target == DNN_TARGET_CPU ? CV_TEST_TAG_MEMORY_512MB : CV_TEST_TAG_MEMORY_2GB));
applyTestTag(
(target == DNN_TARGET_CPU ? CV_TEST_TAG_MEMORY_512MB : CV_TEST_TAG_MEMORY_2GB),
CV_TEST_TAG_LONG,
CV_TEST_TAG_DEBUG_VERYLONG
);
if ((backend == DNN_BACKEND_INFERENCE_ENGINE && target == DNN_TARGET_OPENCL_FP16) ||
(backend == DNN_BACKEND_INFERENCE_ENGINE && target == DNN_TARGET_MYRIAD))
throw SkipTestException("");
modules/dnn/test/test_onnx_importer.cpp
+5 -2
View File
@@ -343,7 +343,7 @@ TEST_P(Test_ONNX_nets, VGG16_bn)
TEST_P(Test_ONNX_nets, ZFNet)
{
applyTestTag(target == DNN_TARGET_CPU ? CV_TEST_TAG_MEMORY_512MB : CV_TEST_TAG_MEMORY_1GB);
applyTestTag(CV_TEST_TAG_MEMORY_2GB);
testONNXModels("zfnet512", pb);
}
@@ -418,7 +418,10 @@ TEST_P(Test_ONNX_nets, MobileNet_v2)
TEST_P(Test_ONNX_nets, LResNet100E_IR)
{
applyTestTag(target == DNN_TARGET_CPU ? CV_TEST_TAG_MEMORY_512MB : CV_TEST_TAG_MEMORY_1GB);
applyTestTag(
(target == DNN_TARGET_CPU ? CV_TEST_TAG_MEMORY_512MB : CV_TEST_TAG_MEMORY_1GB),
CV_TEST_TAG_DEBUG_LONG
);
if (backend == DNN_BACKEND_INFERENCE_ENGINE &&
(target == DNN_TARGET_OPENCL_FP16 || target == DNN_TARGET_OPENCL || target == DNN_TARGET_MYRIAD))
throw SkipTestException("");
modules/dnn/test/test_tf_importer.cpp
+11 -3
View File
@@ -437,7 +437,12 @@ TEST_P(Test_TensorFlow_nets, MobileNet_v1_SSD)
TEST_P(Test_TensorFlow_nets, Faster_RCNN)
{
applyTestTag(CV_TEST_TAG_LONG, (target == DNN_TARGET_CPU ? CV_TEST_TAG_MEMORY_1GB : CV_TEST_TAG_MEMORY_2GB)); // FIXIT split test
// FIXIT split test
applyTestTag(
(target == DNN_TARGET_CPU ? CV_TEST_TAG_MEMORY_1GB : CV_TEST_TAG_MEMORY_2GB),
CV_TEST_TAG_LONG,
CV_TEST_TAG_DEBUG_VERYLONG
);
static std::string names[] = {"faster_rcnn_inception_v2_coco_2018_01_28",
"faster_rcnn_resnet50_coco_2018_01_28"};
@@ -535,7 +540,10 @@ TEST_P(Test_TensorFlow_nets, opencv_face_detector_uint8)
// np.save('east_text_detection.geometry.npy', geometry)
TEST_P(Test_TensorFlow_nets, EAST_text_detection)
{
applyTestTag(target == DNN_TARGET_CPU ? CV_TEST_TAG_MEMORY_512MB : CV_TEST_TAG_MEMORY_1GB);
applyTestTag(
(target == DNN_TARGET_CPU ? CV_TEST_TAG_MEMORY_512MB : CV_TEST_TAG_MEMORY_1GB),
CV_TEST_TAG_DEBUG_LONG
);
#if defined(INF_ENGINE_RELEASE)
if (backend == DNN_BACKEND_INFERENCE_ENGINE && target == DNN_TARGET_MYRIAD)
@@ -765,7 +773,7 @@ TEST(Test_TensorFlow, two_inputs)
TEST(Test_TensorFlow, Mask_RCNN)
{
applyTestTag(CV_TEST_TAG_MEMORY_1GB);
applyTestTag(CV_TEST_TAG_MEMORY_1GB, CV_TEST_TAG_DEBUG_VERYLONG);
std::string proto = findDataFile("dnn/mask_rcnn_inception_v2_coco_2018_01_28.pbtxt", false);
std::string model = findDataFile("dnn/mask_rcnn_inception_v2_coco_2018_01_28.pb", false);