mirror/opencv
1
0
Fork 0
mirror of https://github.com/opencv/opencv.git synced 2024-11-27 20:50:25 +08:00
Code Issues Actions Packages Projects Releases Wiki Activity

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

Browse Source
This commit is contained in:
Alexander Alekhin 2020-02-20 19:45:21 +03:00
commit 96b26dc8f4
28 changed files with 1115 additions and 147 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
apps/interactive-calibration/calibCommon.hpp
View File

@ -80,7 +80,7 @@ namespace calib
cv::Size boardSize;
int charucoDictName;
int calibrationStep;
float charucoSquareLenght, charucoMarkerSize;
float charucoSquareLength, charucoMarkerSize;
float captureDelay;
float squareSize;
float templDst;

2
apps/interactive-calibration/defaultConfig.xml
View File

@ -1,7 +1,7 @@
<?xml version="1.0"?>
<opencv_storage>
<charuco_dict>0</charuco_dict>
<charuco_square_lenght>200</charuco_square_lenght>
<charuco_square_length>200</charuco_square_length>
<charuco_marker_size>100</charuco_marker_size>
<calibration_step>1</calibration_step>
<max_frames_num>30</max_frames_num>

2
apps/interactive-calibration/frameProcessor.cpp
View File

@ -273,7 +273,7 @@ CalibProcessor::CalibProcessor(cv::Ptr<calibrationData> data, captureParameters
#ifdef HAVE_OPENCV_ARUCO
mArucoDictionary = cv::aruco::getPredefinedDictionary(
cv::aruco::PREDEFINED_DICTIONARY_NAME(capParams.charucoDictName));
mCharucoBoard = cv::aruco::CharucoBoard::create(mBoardSize.width, mBoardSize.height, capParams.charucoSquareLenght,
mCharucoBoard = cv::aruco::CharucoBoard::create(mBoardSize.width, mBoardSize.height, capParams.charucoSquareLength,
capParams.charucoMarkerSize, mArucoDictionary);
#endif
break;

2
apps/interactive-calibration/main.cpp
View File

@ -181,7 +181,7 @@ int main(int argc, char** argv)
cv::aruco::getPredefinedDictionary(cv::aruco::PREDEFINED_DICTIONARY_NAME(capParams.charucoDictName));
cv::Ptr<cv::aruco::CharucoBoard> charucoboard =
cv::aruco::CharucoBoard::create(capParams.boardSize.width, capParams.boardSize.height,
capParams.charucoSquareLenght, capParams.charucoMarkerSize, dictionary);
capParams.charucoSquareLength, capParams.charucoMarkerSize, dictionary);
globalData->totalAvgErr =
cv::aruco::calibrateCameraCharuco(globalData->allCharucoCorners, globalData->allCharucoIds,
charucoboard, globalData->imageSize,

9
apps/interactive-calibration/parametersController.cpp
View File

@ -37,7 +37,10 @@ bool calib::parametersController::loadFromFile(const std::string &inputFileName)
}
readFromNode(reader["charuco_dict"], mCapParams.charucoDictName);
readFromNode(reader["charuco_square_lenght"], mCapParams.charucoSquareLenght);
if (readFromNode(reader["charuco_square_lenght"], mCapParams.charucoSquareLength)) {
std::cout << "DEPRECATION: Parameter 'charuco_square_lenght' has been deprecated (typo). Use 'charuco_square_length' instead." << std::endl;
}
readFromNode(reader["charuco_square_length"], mCapParams.charucoSquareLength);
readFromNode(reader["charuco_marker_size"], mCapParams.charucoMarkerSize);
readFromNode(reader["camera_resolution"], mCapParams.cameraResolution);
readFromNode(reader["calibration_step"], mCapParams.calibrationStep);
@ -51,7 +54,7 @@ bool calib::parametersController::loadFromFile(const std::string &inputFileName)
bool retValue =
checkAssertion(mCapParams.charucoDictName >= 0, "Dict name must be >= 0") &&
checkAssertion(mCapParams.charucoMarkerSize > 0, "Marker size must be positive") &&
checkAssertion(mCapParams.charucoSquareLenght > 0, "Square size must be positive") &&
checkAssertion(mCapParams.charucoSquareLength > 0, "Square size must be positive") &&
checkAssertion(mCapParams.minFramesNum > 1, "Minimal number of frames for calibration < 1") &&
checkAssertion(mCapParams.calibrationStep > 0, "Calibration step must be positive") &&
checkAssertion(mCapParams.maxFramesNum > mCapParams.minFramesNum, "maxFramesNum < minFramesNum") &&
@ -119,7 +122,7 @@ bool calib::parametersController::loadFromParser(cv::CommandLineParser &parser)
mCapParams.board = chAruco;
mCapParams.boardSize = cv::Size(6, 8);
mCapParams.charucoDictName = 0;
mCapParams.charucoSquareLenght = 200;
mCapParams.charucoSquareLength = 200;
mCapParams.charucoMarkerSize = 100;
}
else {

6
doc/tutorials/calib3d/interactive_calibration/interactive_calibration.markdown
View File

@ -64,7 +64,7 @@ By default values of advanced parameters are stored in defaultConfig.xml
<?xml version="1.0"?>
<opencv_storage>
<charuco_dict>0</charuco_dict>
<charuco_square_lenght>200</charuco_square_lenght>
<charuco_square_length>200</charuco_square_length>
<charuco_marker_size>100</charuco_marker_size>
<calibration_step>1</calibration_step>
<max_frames_num>30</max_frames_num>
@ -78,7 +78,7 @@ By default values of advanced parameters are stored in defaultConfig.xml
@endcode
- *charuco_dict*: name of special dictionary, which has been used for generation of chAruco pattern
- *charuco_square_lenght*: size of square on chAruco board (in pixels)
- *charuco_square_length*: size of square on chAruco board (in pixels)
- *charuco_marker_size*: size of Aruco markers on chAruco board (in pixels)
- *calibration_step*: interval in frames between launches of @ref cv::calibrateCamera
- *max_frames_num*: if number of frames for calibration is greater then this value frames filter starts working.
@ -91,7 +91,7 @@ QR faster than SVD, but potentially less precise
- *frame_filter_conv_param*: parameter which used in linear convolution of bicriterial frames filter
- *camera_resolution*: resolution of camera which is used for calibration
**Note:** *charuco_dict*, *charuco_square_lenght* and *charuco_marker_size* are used for chAruco pattern generation
**Note:** *charuco_dict*, *charuco_square_length* and *charuco_marker_size* are used for chAruco pattern generation
(see Aruco module description for details: [Aruco tutorials](https://github.com/opencv/opencv_contrib/tree/master/modules/aruco/tutorials))
Default chAruco pattern:

51
modules/calib3d/src/fundam.cpp
View File

@ -490,12 +490,47 @@ static int run7Point( const Mat& _m1, const Mat& _m2, Mat& _fmatrix )
double* fmatrix = _fmatrix.ptr<double>();
int i, k, n;
Point2d m1c(0, 0), m2c(0, 0);
double t, scale1 = 0, scale2 = 0;
const int count = 7;
// compute centers and average distances for each of the two point sets
for( i = 0; i < count; i++ )
{
m1c += Point2d(m1[i]);
m2c += Point2d(m2[i]);
}
// calculate the normalizing transformations for each of the point sets:
// after the transformation each set will have the mass center at the coordinate origin
// and the average distance from the origin will be ~sqrt(2).
t = 1./count;
m1c *= t;
m2c *= t;
for( i = 0; i < count; i++ )
{
scale1 += norm(Point2d(m1[i].x - m1c.x, m1[i].y - m1c.y));
scale2 += norm(Point2d(m2[i].x - m2c.x, m2[i].y - m2c.y));
}
scale1 *= t;
scale2 *= t;
if( scale1 < FLT_EPSILON || scale2 < FLT_EPSILON )
return 0;
scale1 = std::sqrt(2.)/scale1;
scale2 = std::sqrt(2.)/scale2;
// form a linear system: i-th row of A(=a) represents
// the equation: (m2[i], 1)'*F*(m1[i], 1) = 0
for( i = 0; i < 7; i++ )
{
double x0 = m1[i].x, y0 = m1[i].y;
double x1 = m2[i].x, y1 = m2[i].y;
double x0 = (m1[i].x - m1c.x)*scale1;
double y0 = (m1[i].y - m1c.y)*scale1;
double x1 = (m2[i].x - m2c.x)*scale2;
double y1 = (m2[i].y - m2c.y)*scale2;
a[i*9+0] = x1*x0;
a[i*9+1] = x1*y0;
@ -559,6 +594,10 @@ static int run7Point( const Mat& _m1, const Mat& _m2, Mat& _fmatrix )
if( n < 1 || n > 3 )
return n;
// transformation matrices
Matx33d T1( scale1, 0, -scale1*m1c.x, 0, scale1, -scale1*m1c.y, 0, 0, 1 );
Matx33d T2( scale2, 0, -scale2*m2c.x, 0, scale2, -scale2*m2c.y, 0, 0, 1 );
for( k = 0; k < n; k++, fmatrix += 9 )
{
// for each root form the fundamental matrix
@ -577,6 +616,14 @@ static int run7Point( const Mat& _m1, const Mat& _m2, Mat& _fmatrix )
for( i = 0; i < 8; i++ )
fmatrix[i] = f1[i]*lambda + f2[i]*mu;
// de-normalize
Mat F(3, 3, CV_64F, fmatrix);
F = T2.t() * F * T1;
// make F(3,3) = 1
if(fabs(F.at<double>(8)) > FLT_EPSILON )
F *= 1. / F.at<double>(8);
}
return n;

8
modules/core/src/persistence_yml.cpp
View File

@ -452,19 +452,19 @@ public:
if ( d == '<') //support of full type heading from YAML 1.2
{
const char* yamlTypeHeading = "<tag:yaml.org,2002:";
const size_t headingLenght = strlen(yamlTypeHeading);
const size_t headingLength = strlen(yamlTypeHeading);
char* typeEndPtr = ++ptr;
do d = *++typeEndPtr;
while( cv_isprint(d) && d != ' ' && d != '>' );
if ( d == '>' && (size_t)(typeEndPtr - ptr) > headingLenght )
if ( d == '>' && (size_t)(typeEndPtr - ptr) > headingLength )
{
if ( memcmp(ptr, yamlTypeHeading, headingLenght) == 0 )
if ( memcmp(ptr, yamlTypeHeading, headingLength) == 0 )
{
*typeEndPtr = ' ';
ptr += headingLenght - 1;
ptr += headingLength - 1;
is_user = true;
//value_type |= FileNode::USER;
}

98
modules/dnn/src/caffe/caffe_importer.cpp
View File

@ -75,6 +75,17 @@ static cv::String toString(const T &v)
return ss.str();
}
static inline
MatShape parseBlobShape(const caffe::BlobShape& _input_shape)
{
MatShape shape;
for (int i = 0; i < _input_shape.dim_size(); i++)
{
shape.push_back((int)_input_shape.dim(i));
}
return shape;
}
class CaffeImporter
{
caffe::NetParameter net;
@ -235,10 +246,7 @@ public:
}
else if (pbBlob.has_shape())
{
const caffe::BlobShape &_shape = pbBlob.shape();
for (int i = 0; i < _shape.dim_size(); i++)
shape.push_back((int)_shape.dim(i));
shape = parseBlobShape(pbBlob.shape());
}
else
shape.resize(1, 1); // Is a scalar.
@ -334,12 +342,49 @@ public:
//setup input layer names
std::vector<String> netInputs(net.input_size());
std::vector<MatShape> inp_shapes;
{
for (int inNum = 0; inNum < net.input_size(); inNum++)
int net_input_size = net.input_size();
for (int inNum = 0; inNum < net_input_size; inNum++)
{
addedBlobs.push_back(BlobNote(net.input(inNum), 0, inNum));
netInputs[inNum] = net.input(inNum);
}
if (net.input_dim_size() > 0) // deprecated in Caffe proto
{
int net_input_dim_size = net.input_dim_size();
CV_Check(net_input_dim_size, net_input_dim_size % 4 == 0, "");
CV_CheckEQ(net_input_dim_size, net_input_size * 4, "");
for (int inp_id = 0; inp_id < net_input_size; inp_id++)
{
int dim = inp_id * 4;
MatShape shape(4);
shape[0] = net.input_dim(dim);
shape[1] = net.input_dim(dim+1);
shape[2] = net.input_dim(dim+2);
shape[3] = net.input_dim(dim+3);
inp_shapes.push_back(shape);
}
}
else if (net.input_shape_size() > 0) // deprecated in Caffe proto
{
int net_input_shape_size = net.input_shape_size();
CV_CheckEQ(net_input_shape_size, net_input_size, "");
for (int inp_id = 0; inp_id < net_input_shape_size; inp_id++)
{
MatShape shape = parseBlobShape(net.input_shape(inp_id));
inp_shapes.push_back(shape);
}
}
else
{
for (int inp_id = 0; inp_id < net_input_size; inp_id++)
{
MatShape shape; // empty
inp_shapes.push_back(shape);
}
}
}
for (int li = 0; li < layersSize; li++)
@ -364,6 +409,17 @@ public:
addedBlobs.back().outNum = netInputs.size();
netInputs.push_back(addedBlobs.back().name);
}
if (layer.has_input_param())
{
const caffe::InputParameter &inputParameter = layer.input_param();
int input_shape_size = inputParameter.shape_size();
CV_CheckEQ(input_shape_size, layer.top_size(), "");
for (int inp_id = 0; inp_id < input_shape_size; inp_id++)
{
MatShape shape = parseBlobShape(inputParameter.shape(inp_id));
inp_shapes.push_back(shape);
}
}
continue;
}
else if (type == "BatchNorm")
@ -424,35 +480,15 @@ public:
}
dstNet.setInputsNames(netInputs);
std::vector<MatShape> inp_shapes;
if (net.input_shape_size() > 0 || (layersSize > 0 && net.layer(0).has_input_param() &&
net.layer(0).input_param().shape_size() > 0)) {
int size = (net.input_shape_size() > 0) ? net.input_shape_size() :
net.layer(0).input_param().shape_size();
for (int inp_id = 0; inp_id < size; inp_id++)
if (inp_shapes.size() > 0)
{
CV_CheckEQ(inp_shapes.size(), netInputs.size(), "");
for (int inp_id = 0; inp_id < inp_shapes.size(); inp_id++)
{
const caffe::BlobShape &_input_shape = (net.input_shape_size() > 0) ?
net.input_shape(inp_id) :
net.layer(0).input_param().shape(inp_id);
MatShape shape;
for (int i = 0; i < _input_shape.dim_size(); i++) {
shape.push_back((int)_input_shape.dim(i));
}
inp_shapes.push_back(shape);
if (!inp_shapes[inp_id].empty())
dstNet.setInput(Mat(inp_shapes[inp_id], CV_32F), netInputs[inp_id]);
}
}
else if (net.input_dim_size() > 0) {
MatShape shape;
for (int dim = 0; dim < net.input_dim_size(); dim++) {
shape.push_back(net.input_dim(dim));
}
inp_shapes.push_back(shape);
}
for (int inp_id = 0; inp_id < inp_shapes.size(); inp_id++) {
dstNet.setInput(Mat(inp_shapes[inp_id], CV_32F), netInputs[inp_id]);
}
addedBlobs.clear();
}

32
modules/dnn/src/dnn.cpp
View File

@ -1418,13 +1418,15 @@ struct Net::Impl
clear();
this->blobsToKeep = blobsToKeep_;
allocateLayers(blobsToKeep_);
MapIdToLayerData::iterator it = layers.find(0);
CV_Assert(it != layers.end());
it->second.skip = netInputLayer->skip;
initBackend();
initBackend(blobsToKeep_);
if (!netWasAllocated)
{
@ -1437,7 +1439,6 @@ struct Net::Impl
}
netWasAllocated = true;
this->blobsToKeep = blobsToKeep_;
if (DNN_NETWORK_DUMP > 0)
{
@ -1564,7 +1565,7 @@ struct Net::Impl
ldOut.consumers.push_back(LayerPin(inLayerId, outNum));
}
void initBackend()
void initBackend(const std::vector<LayerPin>& blobsToKeep_)
{
CV_TRACE_FUNCTION();
if (preferableBackend == DNN_BACKEND_OPENCV)
@ -1574,7 +1575,7 @@ struct Net::Impl
else if (preferableBackend == DNN_BACKEND_INFERENCE_ENGINE_NN_BUILDER_2019)
{
#ifdef HAVE_INF_ENGINE
initInfEngineBackend();
initInfEngineBackend(blobsToKeep_);
#else
CV_Assert(false && "This OpenCV version is built without Inference Engine API support");
#endif
@ -1582,7 +1583,7 @@ struct Net::Impl
else if (preferableBackend == DNN_BACKEND_INFERENCE_ENGINE_NGRAPH)
{
#ifdef HAVE_DNN_NGRAPH
initNgraphBackend();
initNgraphBackend(blobsToKeep_);
#else
CV_Error(Error::StsNotImplemented, "This OpenCV version is built without support of Inference Engine + nGraph");
#endif
@ -1688,7 +1689,7 @@ struct Net::Impl
}
}
void initInfEngineBackend()
void initInfEngineBackend(const std::vector<LayerPin>& blobsToKeep_)
{
CV_TRACE_FUNCTION();
CV_Assert_N(preferableBackend == DNN_BACKEND_INFERENCE_ENGINE_NN_BUILDER_2019, haveInfEngine());
@ -1878,6 +1879,15 @@ struct Net::Impl
CV_Assert(!ieNode.empty());
ieNode->net = net;
for (const auto& pin : blobsToKeep_)
{
if (pin.lid == ld.id)
{
ieNode->net->addOutput(ieNode->layer.getName());
break;
}
}
// Convert weights in FP16 for specific targets.
if ((preferableTarget == DNN_TARGET_OPENCL_FP16 ||
preferableTarget == DNN_TARGET_MYRIAD ||
@ -1984,7 +1994,7 @@ struct Net::Impl
}
}
void initNgraphBackend()
void initNgraphBackend(const std::vector<LayerPin>& blobsToKeep_)
{
CV_TRACE_FUNCTION();
CV_Assert_N(preferableBackend == DNN_BACKEND_INFERENCE_ENGINE_NGRAPH, haveInfEngine());
@ -2173,6 +2183,14 @@ struct Net::Impl
// TF EAST_text_detection
ieNode->net->setUnconnectedNodes(ieNode);
}
for (const auto& pin : blobsToKeep_)
{
if (pin.lid == ld.id)
{
ieNode->net->addOutput(ieNode->node->get_friendly_name());
break;
}
}
ieNode->net->setNodePtr(&ieNode->node);
net->addBlobs(ld.inputBlobsWrappers);

9
modules/dnn/src/ie_ngraph.cpp
View File

@ -231,11 +231,10 @@ void InfEngineNgraphNet::init(Target targetId)
}
}
}
} else {
for (const auto& name : requestedOutputs)
{
cnn.addOutput(name);
}
}
for (const auto& name : requestedOutputs)
{
cnn.addOutput(name);
}
for (const auto& it : cnn.getInputsInfo())

49
modules/dnn/src/layers/crop_and_resize_layer.cpp
View File

@ -5,6 +5,7 @@
// Copyright (C) 2018, Intel Corporation, all rights reserved.
// Third party copyrights are property of their respective owners.
#include "../precomp.hpp"
#include "../ie_ngraph.hpp"
#include "layers_common.hpp"
#ifdef HAVE_CUDA
@ -25,6 +26,14 @@ public:
outHeight = params.get<float>("height");
}
virtual bool supportBackend(int backendId) CV_OVERRIDE
{
return backendId == DNN_BACKEND_OPENCV
|| backendId == DNN_BACKEND_INFERENCE_ENGINE_NGRAPH
|| backendId == DNN_BACKEND_CUDA
;
}
bool getMemoryShapes(const std::vector<MatShape> &inputs,
const int requiredOutputs,
std::vector<MatShape> &outputs,
@ -41,11 +50,6 @@ public:
return false;
}
virtual bool supportBackend(int backendId) CV_OVERRIDE
{
return backendId == DNN_BACKEND_OPENCV || backendId == DNN_BACKEND_CUDA;
}
void forward(InputArrayOfArrays inputs_arr, OutputArrayOfArrays outputs_arr, OutputArrayOfArrays internals_arr) CV_OVERRIDE
{
CV_TRACE_FUNCTION();
@ -121,6 +125,41 @@ public:
}
}
#ifdef HAVE_DNN_NGRAPH
virtual Ptr<BackendNode> initNgraph(const std::vector<Ptr<BackendWrapper> >& inputs,
const std::vector<Ptr<BackendNode> >& nodes) CV_OVERRIDE
{
// Slice second input: from 1x1xNx7 to 1x1xNx5
auto input = nodes[0].dynamicCast<InfEngineNgraphNode>()->node;
auto rois = nodes[1].dynamicCast<InfEngineNgraphNode>()->node;
std::vector<size_t> dims = rois->get_shape(), offsets(4, 0);
offsets[3] = 2;
dims[3] = 7;
auto lower_bounds = std::make_shared<ngraph::op::Constant>(ngraph::element::i64,
ngraph::Shape{offsets.size()}, offsets.data());
auto upper_bounds = std::make_shared<ngraph::op::Constant>(ngraph::element::i64,
ngraph::Shape{dims.size()}, dims.data());
auto strides = std::make_shared<ngraph::op::Constant>(ngraph::element::i64,
ngraph::Shape{dims.size()}, std::vector<int64_t>((int64_t)dims.size(), 1));
auto slice = std::make_shared<ngraph::op::v1::StridedSlice>(rois,
lower_bounds, upper_bounds, strides, std::vector<int64_t>{}, std::vector<int64_t>{});
// Reshape rois from 4D to 2D
std::vector<size_t> shapeData = {dims[2], 5};
auto shape = std::make_shared<ngraph::op::Constant>(ngraph::element::i64, ngraph::Shape{2}, shapeData.data());
auto reshape = std::make_shared<ngraph::op::v1::Reshape>(slice, shape, true);
auto roiPooling =
std::make_shared<ngraph::op::v0::ROIPooling>(input, reshape,
ngraph::Shape{(size_t)outHeight, (size_t)outWidth},
1.0f, "bilinear");
return Ptr<BackendNode>(new InfEngineNgraphNode(roiPooling));
}
#endif // HAVE_DNN_NGRAPH
#ifdef HAVE_CUDA
Ptr<BackendNode> initCUDA(
void *context_,

7
modules/dnn/src/layers/resize_layer.cpp
View File

@ -41,7 +41,7 @@ public:
CV_Assert(params.has("zoom_factor_x") && params.has("zoom_factor_y"));
}
interpolation = params.get<String>("interpolation");
CV_Assert(interpolation == "nearest" || interpolation == "bilinear");
CV_Assert(interpolation == "nearest" || interpolation == "opencv_linear" || interpolation == "bilinear");
alignCorners = params.get<bool>("align_corners", false);
}
@ -115,14 +115,15 @@ public:
Mat& inp = inputs[0];
Mat& out = outputs[0];
if (interpolation == "nearest")
if (interpolation == "nearest" || interpolation == "opencv_linear")
{
InterpolationFlags mode = interpolation == "nearest" ? INTER_NEAREST : INTER_LINEAR;
for (size_t n = 0; n < inputs[0].size[0]; ++n)
{
for (size_t ch = 0; ch < inputs[0].size[1]; ++ch)
{
resize(getPlane(inp, n, ch), getPlane(out, n, ch),
Size(outWidth, outHeight), 0, 0, INTER_NEAREST);
Size(outWidth, outHeight), 0, 0, mode);
}
}
}

37
modules/dnn/src/layers/scale_layer.cpp
View File

@ -61,7 +61,8 @@ public:
return backendId == DNN_BACKEND_OPENCV ||
backendId == DNN_BACKEND_CUDA ||
backendId == DNN_BACKEND_HALIDE ||
((backendId == DNN_BACKEND_INFERENCE_ENGINE_NN_BUILDER_2019 || backendId == DNN_BACKEND_INFERENCE_ENGINE_NGRAPH) && axis == 1);
(backendId == DNN_BACKEND_INFERENCE_ENGINE_NN_BUILDER_2019 && axis == 1) ||
(backendId == DNN_BACKEND_INFERENCE_ENGINE_NGRAPH && axis > 0);
}
void forward(InputArrayOfArrays inputs_arr, OutputArrayOfArrays outputs_arr, OutputArrayOfArrays internals_arr) CV_OVERRIDE
@ -263,22 +264,26 @@ public:
auto ieInpNode = nodes[0].dynamicCast<InfEngineNgraphNode>()->node;
std::vector<size_t> shape(ieInpNode->get_shape().size(), 1);
shape[1] = numChannels;
auto weight = hasWeights ?
std::make_shared<ngraph::op::Constant>(ngraph::element::f32,
ngraph::Shape(shape), blobs[0].data) :
std::make_shared<ngraph::op::Constant>(ngraph::element::f32,
ngraph::Shape(shape), std::vector<float>(numChannels, 1).data());
int cAxis = clamp(axis, shape.size());
shape[cAxis] = numChannels;
auto bias = hasBias ?
std::make_shared<ngraph::op::Constant>(ngraph::element::f32,
ngraph::Shape(shape), blobs.back().data) :
std::make_shared<ngraph::op::Constant>(ngraph::element::f32,
ngraph::Shape(shape), std::vector<float>(numChannels, 0).data());
auto scale_node = std::make_shared<ngraph::op::v1::Multiply>(ieInpNode, weight, ngraph::op::AutoBroadcastType::NUMPY);
auto scale_shift = std::make_shared<ngraph::op::v1::Add>(scale_node, bias, ngraph::op::AutoBroadcastType::NUMPY);
return Ptr<BackendNode>(new InfEngineNgraphNode(scale_shift));
auto node = ieInpNode;
if (hasWeights)
{
auto weight = std::make_shared<ngraph::op::Constant>(ngraph::element::f32,
ngraph::Shape(shape), blobs[0].data);
node = std::make_shared<ngraph::op::v1::Multiply>(node, weight, ngraph::op::AutoBroadcastType::NUMPY);
}
if (hasBias || !hasWeights)
{
auto bias = hasBias ?
std::make_shared<ngraph::op::Constant>(ngraph::element::f32,
ngraph::Shape(shape), blobs.back().data) :
std::make_shared<ngraph::op::Constant>(ngraph::element::f32,
ngraph::Shape(shape), std::vector<float>(numChannels, 0).data());
node = std::make_shared<ngraph::op::v1::Add>(node, bias, ngraph::op::AutoBroadcastType::NUMPY);
}
return Ptr<BackendNode>(new InfEngineNgraphNode(node));
}
#endif // HAVE_DNN_NGRAPH

32
modules/dnn/src/onnx/onnx_importer.cpp
View File

@ -485,16 +485,23 @@ void ONNXImporter::populateNet(Net dstNet)
}
else if (layer_type == "Split")
{
DictValue splits = layerParams.get("split");
const int numSplits = splits.size();
CV_Assert(numSplits > 1);
std::vector<int> slicePoints(numSplits - 1, splits.get<int>(0));
for (int i = 1; i < splits.size() - 1; ++i)
if (layerParams.has("split"))
{
slicePoints[i] = slicePoints[i - 1] + splits.get<int>(i - 1);
DictValue splits = layerParams.get("split");
const int numSplits = splits.size();
CV_Assert(numSplits > 1);
std::vector<int> slicePoints(numSplits - 1, splits.get<int>(0));
for (int i = 1; i < splits.size() - 1; ++i)
{
slicePoints[i] = slicePoints[i - 1] + splits.get<int>(i - 1);
}
layerParams.set("slice_point", DictValue::arrayInt(&slicePoints[0], slicePoints.size()));
}
else
{
layerParams.set("num_split", node_proto.output_size());
}
layerParams.set("slice_point", DictValue::arrayInt(&slicePoints[0], slicePoints.size()));
layerParams.type = "Slice";
}
else if (layer_type == "Add" || layer_type == "Sum")
@ -973,6 +980,15 @@ void ONNXImporter::populateNet(Net dstNet)
replaceLayerParam(layerParams, "width_scale", "zoom_factor_x");
}
replaceLayerParam(layerParams, "mode", "interpolation");
if (layerParams.get<String>("interpolation") == "linear" && framework_name == "pytorch") {
layerParams.type = "Resize";
Mat scales = getBlob(node_proto, constBlobs, 1);
CV_Assert(scales.total() == 4);
layerParams.set("interpolation", "opencv_linear");
layerParams.set("zoom_factor_y", scales.at<float>(2));
layerParams.set("zoom_factor_x", scales.at<float>(3));
}
}
else if (layer_type == "LogSoftmax")
{

46
modules/dnn/test/test_ie_models.cpp
View File

@ -73,28 +73,7 @@ struct OpenVINOModelTestCaseInfo
static const std::map<std::string, OpenVINOModelTestCaseInfo>& getOpenVINOTestModels()
{
static std::map<std::string, OpenVINOModelTestCaseInfo> g_models {
#if INF_ENGINE_RELEASE <= 2018050000
{ "age-gender-recognition-retail-0013", {
"deployment_tools/intel_models/age-gender-recognition-retail-0013/FP32/age-gender-recognition-retail-0013",
"deployment_tools/intel_models/age-gender-recognition-retail-0013/FP16/age-gender-recognition-retail-0013"
}},
{ "face-person-detection-retail-0002", {
"deployment_tools/intel_models/face-person-detection-retail-0002/FP32/face-person-detection-retail-0002",
"deployment_tools/intel_models/face-person-detection-retail-0002/FP16/face-person-detection-retail-0002"
}},
{ "head-pose-estimation-adas-0001", {
"deployment_tools/intel_models/head-pose-estimation-adas-0001/FP32/head-pose-estimation-adas-0001",
"deployment_tools/intel_models/head-pose-estimation-adas-0001/FP16/head-pose-estimation-adas-0001"
}},
{ "person-detection-retail-0002", {
"deployment_tools/intel_models/person-detection-retail-0002/FP32/person-detection-retail-0002",
"deployment_tools/intel_models/person-detection-retail-0002/FP16/person-detection-retail-0002"
}},
{ "vehicle-detection-adas-0002", {
"deployment_tools/intel_models/vehicle-detection-adas-0002/FP32/vehicle-detection-adas-0002",
"deployment_tools/intel_models/vehicle-detection-adas-0002/FP16/vehicle-detection-adas-0002"
}}
#else
#if INF_ENGINE_RELEASE >= 2018050000
// layout is defined by open_model_zoo/model_downloader
// Downloaded using these parameters for Open Model Zoo downloader (2019R1):
// ./downloader.py -o ${OPENCV_DNN_TEST_DATA_PATH}/omz_intel_models --cache_dir ${OPENCV_DNN_TEST_DATA_PATH}/.omz_cache/ \
@ -118,7 +97,16 @@ static const std::map<std::string, OpenVINOModelTestCaseInfo>& getOpenVINOTestMo
{ "vehicle-detection-adas-0002", {
"Transportation/object_detection/vehicle/mobilenet-reduced-ssd/dldt/vehicle-detection-adas-0002",
"Transportation/object_detection/vehicle/mobilenet-reduced-ssd/dldt/vehicle-detection-adas-0002-fp16"
}}
}},
#endif
#if INF_ENGINE_RELEASE >= 2020010000
// Downloaded using these parameters for Open Model Zoo downloader (2020.1):
// ./downloader.py -o ${OPENCV_DNN_TEST_DATA_PATH}/omz_intel_models --cache_dir ${OPENCV_DNN_TEST_DATA_PATH}/.omz_cache/ \
// --name person-detection-retail-0013
{ "person-detection-retail-0013", { // IRv10
"intel/person-detection-retail-0013/FP32/person-detection-retail-0013",
"intel/person-detection-retail-0013/FP16/person-detection-retail-0013"
}},
#endif
};
@ -305,8 +293,8 @@ TEST_P(DNNTestOpenVINO, models)
OpenVINOModelTestCaseInfo modelInfo = it->second;
std::string modelPath = isFP16 ? modelInfo.modelPathFP16 : modelInfo.modelPathFP32;
std::string xmlPath = findDataFile(modelPath + ".xml");
std::string binPath = findDataFile(modelPath + ".bin");
std::string xmlPath = findDataFile(modelPath + ".xml", false);
std::string binPath = findDataFile(modelPath + ".bin", false);
std::map<std::string, cv::Mat> inputsMap;
std::map<std::string, cv::Mat> ieOutputsMap, cvOutputsMap;
@ -316,13 +304,19 @@ TEST_P(DNNTestOpenVINO, models)
runIE(targetId, xmlPath, binPath, inputsMap, ieOutputsMap);
runCV(backendId, targetId, xmlPath, binPath, inputsMap, cvOutputsMap);
double eps = 0;
#if INF_ENGINE_VER_MAJOR_GE(2020010000)
if (targetId == DNN_TARGET_CPU && checkHardwareSupport(CV_CPU_AVX_512F))
eps = 1e-5;
#endif
EXPECT_EQ(ieOutputsMap.size(), cvOutputsMap.size());
for (auto& srcIt : ieOutputsMap)
{
auto dstIt = cvOutputsMap.find(srcIt.first);
CV_Assert(dstIt != cvOutputsMap.end());
double normInf = cvtest::norm(srcIt.second, dstIt->second, cv::NORM_INF);
EXPECT_EQ(normInf, 0);
EXPECT_LE(normInf, eps) << "output=" << srcIt.first;
}
}

15
modules/dnn/test/test_onnx_importer.cpp
View File

@ -335,6 +335,9 @@ TEST_P(Test_ONNX_layers, Padding)
TEST_P(Test_ONNX_layers, Resize)
{
testONNXModels("resize_nearest");
if (backend == DNN_BACKEND_INFERENCE_ENGINE_NN_BUILDER_2019)
applyTestTag(CV_TEST_TAG_DNN_SKIP_IE_NN_BUILDER);
testONNXModels("resize_bilinear");
}
TEST_P(Test_ONNX_layers, MultyInputs)
@ -411,6 +414,18 @@ TEST_P(Test_ONNX_layers, ReduceL2)
testONNXModels("reduceL2");
}
TEST_P(Test_ONNX_layers, Split)
{
if (backend == DNN_BACKEND_INFERENCE_ENGINE_NN_BUILDER_2019)
applyTestTag(CV_TEST_TAG_DNN_SKIP_IE_NN_BUILDER);
if (backend == DNN_BACKEND_INFERENCE_ENGINE_NGRAPH)
applyTestTag(CV_TEST_TAG_DNN_SKIP_IE_NGRAPH);
testONNXModels("split_1");
testONNXModels("split_2");
testONNXModels("split_3");
testONNXModels("split_4");
}
TEST_P(Test_ONNX_layers, Slice)
{
#if defined(INF_ENGINE_RELEASE) && INF_ENGINE_VER_MAJOR_LT(2019010000)

31
modules/dnn/test/test_tf_importer.cpp
View File

@ -994,8 +994,16 @@ TEST(Test_TensorFlow, two_inputs)
normAssert(out, firstInput + secondInput);
}
TEST(Test_TensorFlow, Mask_RCNN)
TEST_P(Test_TensorFlow_nets, Mask_RCNN)
{
static const double kMaskThreshold = 0.5;
if (backend == DNN_BACKEND_INFERENCE_ENGINE_NN_BUILDER_2019 && target == DNN_TARGET_MYRIAD)
applyTestTag(CV_TEST_TAG_DNN_SKIP_IE_MYRIAD, CV_TEST_TAG_DNN_SKIP_IE_NN_BUILDER);
if (target == DNN_TARGET_MYRIAD && getInferenceEngineVPUType() == CV_DNN_INFERENCE_ENGINE_VPU_TYPE_MYRIAD_X)
applyTestTag(CV_TEST_TAG_DNN_SKIP_IE_MYRIAD_X);
applyTestTag(CV_TEST_TAG_MEMORY_1GB, CV_TEST_TAG_DEBUG_VERYLONG);
Mat img = imread(findDataFile("dnn/street.png"));
std::string proto = findDataFile("dnn/mask_rcnn_inception_v2_coco_2018_01_28.pbtxt");
@ -1006,7 +1014,8 @@ TEST(Test_TensorFlow, Mask_RCNN)
Mat refMasks = blobFromNPY(path("mask_rcnn_inception_v2_coco_2018_01_28.detection_masks.npy"));
Mat blob = blobFromImage(img, 1.0f, Size(800, 800), Scalar(), true, false);
net.setPreferableBackend(DNN_BACKEND_OPENCV);
net.setPreferableBackend(backend);
net.setPreferableTarget(target);
net.setInput(blob);
@ -1020,7 +1029,10 @@ TEST(Test_TensorFlow, Mask_RCNN)
Mat outDetections = outs[0];
Mat outMasks = outs[1];
normAssertDetections(refDetections, outDetections, "", /*threshold for zero confidence*/1e-5);
double scoreDiff = (target == DNN_TARGET_OPENCL_FP16 || target == DNN_TARGET_MYRIAD) ? 0.019 : 2e-5;
double iouDiff = (target == DNN_TARGET_OPENCL_FP16 || target == DNN_TARGET_MYRIAD) ? 0.018 : default_lInf;
normAssertDetections(refDetections, outDetections, "", /*threshold for zero confidence*/1e-5, scoreDiff, iouDiff);
// Output size of masks is NxCxHxW where
// N - number of detected boxes
@ -1044,7 +1056,18 @@ TEST(Test_TensorFlow, Mask_RCNN)
outMasks(srcRanges).copyTo(masks(dstRanges));
}
cv::Range topRefMasks[] = {Range::all(), Range(0, numDetections), Range::all(), Range::all()};
normAssert(masks, refMasks(&topRefMasks[0]));
refMasks = refMasks(&topRefMasks[0]);
// make binary masks
cv::threshold(masks.reshape(1, 1), masks, kMaskThreshold, 1, THRESH_BINARY);
cv::threshold(refMasks.reshape(1, 1), refMasks, kMaskThreshold, 1, THRESH_BINARY);
double inter = cv::countNonZero(masks & refMasks);
double area = cv::countNonZero(masks | refMasks);
EXPECT_GE(inter / area, 0.99);
if (backend == DNN_BACKEND_INFERENCE_ENGINE_NGRAPH)
expectNoFallbacks(net);
}
}

28
modules/highgui/src/window_cocoa.mm
View File

@ -112,11 +112,14 @@ static bool wasInitialized = false;
BOOL autosize;
BOOL firstContent;
int status;
int x0, y0;
}
@property(assign) CvMouseCallback mouseCallback;
@property(assign) void *mouseParam;
@property(assign) BOOL autosize;
@property(assign) BOOL firstContent;
@property(assign) int x0;
@property(assign) int y0;
@property(retain) NSMutableDictionary *sliders;
@property(readwrite) int status;
- (CVView *)contentView;
@ -252,6 +255,16 @@ CV_IMPL void cvShowImage( const char* name, const CvArr* arr)
contentSize.height = scaledImageSize.height + [window contentView].sliderHeight;
contentSize.width = std::max<int>(scaledImageSize.width, MIN_SLIDER_WIDTH);
[window setContentSize:contentSize]; //adjust sliders to fit new window size
if([window firstContent])
{
int x = [window x0];
int y = [window y0];
if(x >= 0 && y >= 0)
{
y = [[window screen] visibleFrame].size.height - y;
[window setFrameTopLeftPoint:NSMakePoint(x, y)];
}
}
}
}
[window setFirstContent:NO];
@ -275,7 +288,6 @@ CV_IMPL void cvResizeWindow( const char* name, int width, int height)
CV_IMPL void cvMoveWindow( const char* name, int x, int y)
{
CV_FUNCNAME("cvMoveWindow");
__BEGIN__;
@ -287,8 +299,14 @@ CV_IMPL void cvMoveWindow( const char* name, int x, int y)
//cout << "cvMoveWindow"<< endl;
window = cvGetWindow(name);
if(window) {
y = [[window screen] frame].size.height - y;
[window setFrameTopLeftPoint:NSMakePoint(x, y)];
if([window firstContent]) {
[window setX0:x];
[window setY0:y];
}
else {
y = [[window screen] visibleFrame].size.height - y;
[window setFrameTopLeftPoint:NSMakePoint(x, y)];
}
}
[localpool1 drain];
@ -557,6 +575,8 @@ CV_IMPL int cvNamedWindow( const char* name, int flags )
[window setFrameTopLeftPoint:initContentRect.origin];
[window setFirstContent:YES];
[window setX0:-1];
[window setY0:-1];
[window setContentView:[[CVView alloc] init]];
@ -819,6 +839,8 @@ static NSSize constrainAspectRatio(NSSize base, NSSize constraint) {
@synthesize mouseParam;
@synthesize autosize;
@synthesize firstContent;
@synthesize x0;
@synthesize y0;
@synthesize sliders;
@synthesize status;

15
modules/imgproc/test/test_imgwarp.cpp
View File

@ -1413,6 +1413,21 @@ TEST(Resize, lanczos4_regression_16192)
EXPECT_EQ(cvtest::norm(dst, expected, NORM_INF), 0) << dst(Rect(0,0,8,8));
}
TEST(Resize, DISABLED_nearest_regression_15075) // reverted https://github.com/opencv/opencv/pull/16497
{
const int C = 5;
const int i1 = 5, j1 = 5;
Size src_size(12, 12);
Size dst_size(11, 11);
cv::Mat src = cv::Mat::zeros(src_size, CV_8UC(C)), dst;
for (int j = 0; j < C; j++)
src.col(i1).row(j1).data[j] = 1;
cv::resize(src, dst, dst_size, 0, 0, INTER_NEAREST);
EXPECT_EQ(C, cvtest::norm(dst, NORM_L1)) << src.size;
}
TEST(Imgproc_Warp, multichannel)
{
static const int inter_types[] = {INTER_NEAREST, INTER_AREA, INTER_CUBIC,

2
modules/python/package/template/config.py.in
View File

@ -1,3 +1,5 @@
import os
BINARIES_PATHS = [
@CMAKE_PYTHON_BINARIES_PATH@
] + BINARIES_PATHS

8
modules/python/python_loader.cmake
View File

@ -58,7 +58,13 @@ if(NOT OpenCV_FOUND) # Ignore "standalone" builds of Python bindings
else()
list(APPEND CMAKE_PYTHON_BINARIES_INSTALL_PATH "os.path.join(${CMAKE_PYTHON_EXTENSION_INSTALL_PATH_BASE}, '${OPENCV_LIB_INSTALL_PATH}')")
endif()
string(REPLACE ";" ",\n " CMAKE_PYTHON_BINARIES_PATH "${CMAKE_PYTHON_BINARIES_INSTALL_PATH}")
set(CMAKE_PYTHON_BINARIES_PATH "${CMAKE_PYTHON_BINARIES_INSTALL_PATH}")
if (WIN32 AND HAVE_CUDA)
if (DEFINED CUDA_TOOLKIT_ROOT_DIR)
list(APPEND CMAKE_PYTHON_BINARIES_PATH "os.path.join(os.getenv('CUDA_PATH', '${CUDA_TOOLKIT_ROOT_DIR}'), 'bin')")
endif()
endif()
string(REPLACE ";" ",\n " CMAKE_PYTHON_BINARIES_PATH "${CMAKE_PYTHON_BINARIES_PATH}")
configure_file("${PYTHON_SOURCE_DIR}/package/template/config.py.in" "${__python_loader_install_tmp_path}/cv2/config.py" @ONLY)
install(FILES "${__python_loader_install_tmp_path}/cv2/config.py" DESTINATION "${OPENCV_PYTHON_INSTALL_PATH}/cv2/" COMPONENT python)
endif()

10
samples/cpp/falsecolor.cpp
View File

@ -41,10 +41,10 @@ static Mat DrawMyImage(int thickness,int nbShape)
{
Mat img=Mat::zeros(500,256*thickness+100,CV_8UC1);
int offsetx = 50, offsety = 25;
int lineLenght = 50;
int lineLength = 50;
for (int i=0;i<256;i++)
line(img,Point(thickness*i+ offsetx, offsety),Point(thickness*i+ offsetx, offsety+ lineLenght),Scalar(i), thickness);
line(img,Point(thickness*i+ offsetx, offsety),Point(thickness*i+ offsetx, offsety+ lineLength),Scalar(i), thickness);
RNG r;
Point center;
int radius;
@ -57,19 +57,19 @@ static Mat DrawMyImage(int thickness,int nbShape)
int typeShape = r.uniform(MyCIRCLE, MyELLIPSE+1);
switch (typeShape) {
case MyCIRCLE:
center = Point(r.uniform(offsetx,img.cols- offsetx), r.uniform(offsety + lineLenght, img.rows - offsety));
center = Point(r.uniform(offsetx,img.cols- offsetx), r.uniform(offsety + lineLength, img.rows - offsety));
radius = r.uniform(1, min(offsetx, offsety));
circle(img,center,radius,Scalar(i),-1);
break;
case MyRECTANGLE:
center = Point(r.uniform(offsetx, img.cols - offsetx), r.uniform(offsety + lineLenght, img.rows - offsety));
center = Point(r.uniform(offsetx, img.cols - offsetx), r.uniform(offsety + lineLength, img.rows - offsety));
width = r.uniform(1, min(offsetx, offsety));
height = r.uniform(1, min(offsetx, offsety));
rc = Rect(center-Point(width ,height )/2, center + Point(width , height )/2);
rectangle(img,rc, Scalar(i), -1);
break;
case MyELLIPSE:
center = Point(r.uniform(offsetx, img.cols - offsetx), r.uniform(offsety + lineLenght, img.rows - offsety));
center = Point(r.uniform(offsetx, img.cols - offsetx), r.uniform(offsety + lineLength, img.rows - offsety));
width = r.uniform(1, min(offsetx, offsety));
height = r.uniform(1, min(offsetx, offsety));
angle = r.uniform(0, 180);

20
samples/dnn/human_parsing.py
View File

@ -40,6 +40,7 @@ Follow these steps if you want to convert the original model yourself:
'''
import argparse
import os.path
import numpy as np
import cv2 as cv
@ -48,12 +49,11 @@ backends = (cv.dnn.DNN_BACKEND_DEFAULT, cv.dnn.DNN_BACKEND_INFERENCE_ENGINE, cv.
targets = (cv.dnn.DNN_TARGET_CPU, cv.dnn.DNN_TARGET_OPENCL, cv.dnn.DNN_TARGET_OPENCL_FP16, cv.dnn.DNN_TARGET_MYRIAD)
def preprocess(image_path):
def preprocess(image):
"""
Create 4-dimensional blob from image and flip image
:param image_path: path to input image
:param image: input image
"""
image = cv.imread(image_path)
image_rev = np.flip(image, axis=1)
input = cv.dnn.blobFromImages([image, image_rev], mean=(104.00698793, 116.66876762, 122.67891434))
return input
@ -137,15 +137,15 @@ def decode_labels(gray_image):
return segm
def parse_human(image_path, model_path, backend=cv.dnn.DNN_BACKEND_OPENCV, target=cv.dnn.DNN_TARGET_CPU):
def parse_human(image, model_path, backend=cv.dnn.DNN_BACKEND_OPENCV, target=cv.dnn.DNN_TARGET_CPU):
"""
Prepare input for execution, run net and postprocess output to parse human.
:param image_path: path to input image
:param image: input image
:param model_path: path to JPPNet model
:param backend: name of computation backend
:param target: name of computation target
"""
input = preprocess(image_path)
input = preprocess(image)
input_h, input_w = input.shape[2:]
output = run_net(input, model_path, backend, target)
grayscale_out = postprocess(output, (input_w, input_h))
@ -157,7 +157,7 @@ if __name__ == '__main__':
parser = argparse.ArgumentParser(description='Use this script to run human parsing using JPPNet',
formatter_class=argparse.ArgumentDefaultsHelpFormatter)
parser.add_argument('--input', '-i', required=True, help='Path to input image.')
parser.add_argument('--model', '-m', required=True, help='Path to pb model.')
parser.add_argument('--model', '-m', default='lip_jppnet_384.pb', help='Path to pb model.')
parser.add_argument('--backend', choices=backends, default=cv.dnn.DNN_BACKEND_DEFAULT, type=int,
help="Choose one of computation backends: "
"%d: automatically (by default), "
@ -171,7 +171,11 @@ if __name__ == '__main__':
'%d: VPU' % targets)
args, _ = parser.parse_known_args()
output = parse_human(args.input, args.model, args.backend, args.target)
if not os.path.isfile(args.model):
raise OSError("Model not exist")
image = cv.imread(args.input)
output = parse_human(image, args.model, args.backend, args.target)
winName = 'Deep learning human parsing in OpenCV'
cv.namedWindow(winName, cv.WINDOW_AUTOSIZE)
cv.imshow(winName, output)

465
samples/dnn/virtual_try_on.py Normal file
View File

@ -0,0 +1,465 @@
#!/usr/bin/env python3
'''
You can download the Geometric Matching Module model from https://www.dropbox.com/s/tyhc73xa051grjp/cp_vton_gmm.onnx?dl=0
You can download the Try-On Module model from https://www.dropbox.com/s/q2x97ve2h53j66k/cp_vton_tom.onnx?dl=0
You can download the cloth segmentation model from https://www.dropbox.com/s/qag9vzambhhkvxr/lip_jppnet_384.pb?dl=0
You can find the OpenPose proto in opencv_extra/testdata/dnn/openpose_pose_coco.prototxt
and get .caffemodel using opencv_extra/testdata/dnn/download_models.py
'''
import argparse
import os.path
import numpy as np
import cv2 as cv
from numpy import linalg
from common import findFile
from human_parsing import parse_human
backends = (cv.dnn.DNN_BACKEND_DEFAULT, cv.dnn.DNN_BACKEND_HALIDE, cv.dnn.DNN_BACKEND_INFERENCE_ENGINE, cv.dnn.DNN_BACKEND_OPENCV)
targets = (cv.dnn.DNN_TARGET_CPU, cv.dnn.DNN_TARGET_OPENCL, cv.dnn.DNN_TARGET_OPENCL_FP16, cv.dnn.DNN_TARGET_MYRIAD)
parser = argparse.ArgumentParser(description='Use this script to run virtial try-on using CP-VTON',
formatter_class=argparse.ArgumentDefaultsHelpFormatter)
parser.add_argument('--input_image', '-i', required=True, help='Path to image with person.')
parser.add_argument('--input_cloth', '-c', required=True, help='Path to target cloth image')
parser.add_argument('--gmm_model', '-gmm', default='cp_vton_gmm.onnx', help='Path to Geometric Matching Module .onnx model.')
parser.add_argument('--tom_model', '-tom', default='cp_vton_tom.onnx', help='Path to Try-On Module .onnx model.')
parser.add_argument('--segmentation_model', default='lip_jppnet_384.pb', help='Path to cloth segmentation .pb model.')
parser.add_argument('--openpose_proto', default='openpose_pose_coco.prototxt', help='Path to OpenPose .prototxt model was trained on COCO dataset.')
parser.add_argument('--openpose_model', default='openpose_pose_coco.caffemodel', help='Path to OpenPose .caffemodel model was trained on COCO dataset.')
parser.add_argument('--backend', choices=backends, default=cv.dnn.DNN_BACKEND_DEFAULT, type=int,
help="Choose one of computation backends: "
"%d: automatically (by default), "
"%d: Halide language (http://halide-lang.org/), "
"%d: Intel's Deep Learning Inference Engine (https://software.intel.com/openvino-toolkit), "
"%d: OpenCV implementation" % backends)
parser.add_argument('--target', choices=targets, default=cv.dnn.DNN_TARGET_CPU, type=int,
help='Choose one of target computation devices: '
'%d: CPU target (by default), '
'%d: OpenCL, '
'%d: OpenCL fp16 (half-float precision), '
'%d: VPU' % targets)
args, _ = parser.parse_known_args()
def get_pose_map(image, proto_path, model_path, backend, target, height=256, width=192):
radius = 5
inp = cv.dnn.blobFromImage(image, 1.0 / 255, (width, height))
net = cv.dnn.readNet(proto_path, model_path)
net.setPreferableBackend(backend)
net.setPreferableTarget(target)
net.setInput(inp)
out = net.forward()
threshold = 0.1
_, out_c, out_h, out_w = out.shape
pose_map = np.zeros((height, width, out_c - 1))
# last label: Background
for i in range(0, out.shape[1] - 1):
heatMap = out[0, i, :, :]
keypoint = np.full((height, width), -1)
_, conf, _, point = cv.minMaxLoc(heatMap)
x = width * point[0] // out_w
y = height * point[1] // out_h
if conf > threshold and x > 0 and y > 0:
keypoint[y - radius:y + radius, x - radius:x + radius] = 1
pose_map[:, :, i] = keypoint
pose_map = pose_map.transpose(2, 0, 1)
return pose_map
class BilinearFilter(object):
"""
PIL bilinear resize implementation
image = image.resize((image_width // 16, image_height // 16), Image.BILINEAR)
"""
def _precompute_coeffs(self, inSize, outSize):
filterscale = max(1.0, inSize / outSize)
ksize = int(np.ceil(filterscale)) * 2 + 1
kk = np.zeros(shape=(outSize * ksize, ), dtype=np.float32)
bounds = np.empty(shape=(outSize * 2, ), dtype=np.int32)
centers = (np.arange(outSize) + 0.5) * filterscale + 0.5
bounds[::2] = np.where(centers - filterscale < 0, 0, centers - filterscale)
bounds[1::2] = np.where(centers + filterscale > inSize, inSize, centers + filterscale) - bounds[::2]
xmins = bounds[::2] - centers + 1
points = np.array([np.arange(row) + xmins[i] for i, row in enumerate(bounds[1::2])]) / filterscale
for xx in range(0, outSize):
point = points[xx]
bilinear = np.where(point < 1.0, 1.0 - abs(point), 0.0)
ww = np.sum(bilinear)
kk[xx * ksize : xx * ksize + bilinear.size] = np.where(ww == 0.0, bilinear, bilinear / ww)
return bounds, kk, ksize
def _resample_horizontal(self, out, img, ksize, bounds, kk):
for yy in range(0, out.shape[0]):
for xx in range(0, out.shape[1]):
xmin = bounds[xx * 2 + 0]
xmax = bounds[xx * 2 + 1]
k = kk[xx * ksize : xx * ksize + xmax]
out[yy, xx] = np.round(np.sum(img[yy, xmin : xmin + xmax] * k))
def _resample_vertical(self, out, img, ksize, bounds, kk):
for yy in range(0, out.shape[0]):
ymin = bounds[yy * 2 + 0]
ymax = bounds[yy * 2 + 1]
k = kk[yy * ksize: yy * ksize + ymax]
out[yy] = np.round(np.sum(img[ymin : ymin + ymax, 0:out.shape[1]] * k[:, np.newaxis], axis=0))
def imaging_resample(self, img, xsize, ysize):
height, width, *args = img.shape
bounds_horiz, kk_horiz, ksize_horiz = self._precompute_coeffs(width, xsize)
bounds_vert, kk_vert, ksize_vert = self._precompute_coeffs(height, ysize)
out_hor = np.empty((img.shape[0], xsize), dtype=np.uint8)
self._resample_horizontal(out_hor, img, ksize_horiz, bounds_horiz, kk_horiz)
out = np.empty((ysize, xsize), dtype=np.uint8)
self._resample_vertical(out, out_hor, ksize_vert, bounds_vert, kk_vert)
return out
class CpVton(object):
def __init__(self, gmm_model, tom_model, backend, target):
super(CpVton, self).__init__()
self.gmm_net = cv.dnn.readNet(gmm_model)
self.tom_net = cv.dnn.readNet(tom_model)
self.gmm_net.setPreferableBackend(backend)
self.gmm_net.setPreferableTarget(target)
self.tom_net.setPreferableBackend(backend)
self.tom_net.setPreferableTarget(target)
def prepare_agnostic(self, segm_image, input_image, pose_map, height=256, width=192):
palette = {
'Background' : (0, 0, 0),
'Hat' : (128, 0, 0),
'Hair' : (255, 0, 0),
'Glove' : (0, 85, 0),
'Sunglasses' : (170, 0, 51),
'UpperClothes' : (255, 85, 0),
'Dress' : (0, 0, 85),
'Coat' : (0, 119, 221),
'Socks' : (85, 85, 0),
'Pants' : (0, 85, 85),
'Jumpsuits' : (85, 51, 0),
'Scarf' : (52, 86, 128),
'Skirt' : (0, 128, 0),
'Face' : (0, 0, 255),
'Left-arm' : (51, 170, 221),
'Right-arm' : (0, 255, 255),
'Left-leg' : (85, 255, 170),
'Right-leg' : (170, 255, 85),
'Left-shoe' : (255, 255, 0),
'Right-shoe' : (255, 170, 0)
}
color2label = {val: key for key, val in palette.items()}
head_labels = ['Hat', 'Hair', 'Sunglasses', 'Face', 'Pants', 'Skirt']
segm_image = cv.cvtColor(segm_image, cv.COLOR_BGR2RGB)
phead = np.zeros((1, height, width), dtype=np.float32)
pose_shape = np.zeros((height, width), dtype=np.uint8)
for r in range(height):
for c in range(width):
pixel = tuple(segm_image[r, c])
if tuple(pixel) in color2label:
if color2label[pixel] in head_labels:
phead[0, r, c] = 1
if color2label[pixel] != 'Background':
pose_shape[r, c] = 255
input_image = cv.dnn.blobFromImage(input_image, 1.0 / 127.5, (width, height), mean=(127.5, 127.5, 127.5), swapRB=True)
input_image = input_image.squeeze(0)
img_head = input_image * phead - (1 - phead)
downsample = BilinearFilter()
down = downsample.imaging_resample(pose_shape, width // 16, height // 16)
res_shape = cv.resize(down, (width, height), cv.INTER_LINEAR)
res_shape = cv.dnn.blobFromImage(res_shape, 1.0 / 127.5, mean=(127.5, 127.5, 127.5), swapRB=True)
res_shape = res_shape.squeeze(0)
agnostic = np.concatenate((res_shape, img_head, pose_map), axis=0)
agnostic = np.expand_dims(agnostic, axis=0)
return agnostic
def get_warped_cloth(self, cloth_img, agnostic, height=256, width=192):
cloth = cv.dnn.blobFromImage(cloth_img, 1.0 / 127.5, (width, height), mean=(127.5, 127.5, 127.5), swapRB=True)
self.gmm_net.setInput(agnostic, "input.1")
self.gmm_net.setInput(cloth, "input.18")
theta = self.gmm_net.forward()
grid = self._generate_grid(theta)
warped_cloth = self._bilinear_sampler(cloth, grid).astype(np.float32)
return warped_cloth
def get_tryon(self, agnostic, warp_cloth):
inp = np.concatenate([agnostic, warp_cloth], axis=1)
self.tom_net.setInput(inp)
out = self.tom_net.forward()
p_rendered, m_composite = np.split(out, [3], axis=1)
p_rendered = np.tanh(p_rendered)
m_composite = 1 / (1 + np.exp(-m_composite))
p_tryon = warp_cloth * m_composite + p_rendered * (1 - m_composite)
rgb_p_tryon = cv.cvtColor(p_tryon.squeeze(0).transpose(1, 2, 0), cv.COLOR_BGR2RGB)
rgb_p_tryon = (rgb_p_tryon + 1) / 2
return rgb_p_tryon
def _compute_L_inverse(self, X, Y):
N = X.shape[0]
Xmat = np.tile(X, (1, N))
Ymat = np.tile(Y, (1, N))
P_dist_squared = np.power(Xmat - Xmat.transpose(1, 0), 2) + np.power(Ymat - Ymat.transpose(1, 0), 2)
P_dist_squared[P_dist_squared == 0] = 1
K = np.multiply(P_dist_squared, np.log(P_dist_squared))
O = np.ones([N, 1], dtype=np.float32)
Z = np.zeros([3, 3], dtype=np.float32)
P = np.concatenate([O, X, Y], axis=1)
first = np.concatenate((K, P), axis=1)
second = np.concatenate((P.transpose(1, 0), Z), axis=1)
L = np.concatenate((first, second), axis=0)
Li = linalg.inv(L)
return Li
def _prepare_to_transform(self, out_h=256, out_w=192, grid_size=5):
grid = np.zeros([out_h, out_w, 3], dtype=np.float32)
grid_X, grid_Y = np.meshgrid(np.linspace(-1, 1, out_w), np.linspace(-1, 1, out_h))
grid_X = np.expand_dims(np.expand_dims(grid_X, axis=0), axis=3)
grid_Y = np.expand_dims(np.expand_dims(grid_Y, axis=0), axis=3)
axis_coords = np.linspace(-1, 1, grid_size)
N = grid_size ** 2
P_Y, P_X = np.meshgrid(axis_coords, axis_coords)
P_X = np.reshape(P_X,(-1, 1))
P_Y = np.reshape(P_Y,(-1, 1))
P_X = np.expand_dims(np.expand_dims(np.expand_dims(P_X, axis=2), axis=3), axis=4).transpose(4, 1, 2, 3, 0)
P_Y = np.expand_dims(np.expand_dims(np.expand_dims(P_Y, axis=2), axis=3), axis=4).transpose(4, 1, 2, 3, 0)
return grid_X, grid_Y, N, P_X, P_Y
def _expand_torch(self, X, shape):
if len(X.shape) != len(shape):
return X.flatten().reshape(shape)
else:
axis = [1 if src == dst else dst for src, dst in zip(X.shape, shape)]
return np.tile(X, axis)
def _apply_transformation(self, theta, points, N, P_X, P_Y):
if len(theta.shape) == 2:
theta = np.expand_dims(np.expand_dims(theta, axis=2), axis=3)
batch_size = theta.shape[0]
P_X_base = np.copy(P_X)
P_Y_base = np.copy(P_Y)
Li = self._compute_L_inverse(np.reshape(P_X, (N, -1)), np.reshape(P_Y, (N, -1)))
Li = np.expand_dims(Li, axis=0)
# split theta into point coordinates
Q_X = np.squeeze(theta[:, :N, :, :], axis=3)
Q_Y = np.squeeze(theta[:, N:, :, :], axis=3)
Q_X += self._expand_torch(P_X_base, Q_X.shape)
Q_Y += self._expand_torch(P_Y_base, Q_Y.shape)
points_b = points.shape[0]
points_h = points.shape[1]
points_w = points.shape[2]
P_X = self._expand_torch(P_X, (1, points_h, points_w, 1, N))
P_Y = self._expand_torch(P_Y, (1, points_h, points_w, 1, N))
W_X = self._expand_torch(Li[:,:N,:N], (batch_size, N, N)) @ Q_X
W_Y = self._expand_torch(Li[:,:N,:N], (batch_size, N, N)) @ Q_Y
W_X = np.expand_dims(np.expand_dims(W_X, axis=3), axis=4).transpose(0, 4, 2, 3, 1)
W_X = np.repeat(W_X, points_h, axis=1)
W_X = np.repeat(W_X, points_w, axis=2)
W_Y = np.expand_dims(np.expand_dims(W_Y, axis=3), axis=4).transpose(0, 4, 2, 3, 1)
W_Y = np.repeat(W_Y, points_h, axis=1)
W_Y = np.repeat(W_Y, points_w, axis=2)
A_X = self._expand_torch(Li[:, N:, :N], (batch_size, 3, N)) @ Q_X
A_Y = self._expand_torch(Li[:, N:, :N], (batch_size, 3, N)) @ Q_Y
A_X = np.expand_dims(np.expand_dims(A_X, axis=3), axis=4).transpose(0, 4, 2, 3, 1)
A_X = np.repeat(A_X, points_h, axis=1)
A_X = np.repeat(A_X, points_w, axis=2)
A_Y = np.expand_dims(np.expand_dims(A_Y, axis=3), axis=4).transpose(0, 4, 2, 3, 1)
A_Y = np.repeat(A_Y, points_h, axis=1)
A_Y = np.repeat(A_Y, points_w, axis=2)
points_X_for_summation = np.expand_dims(np.expand_dims(points[:, :, :, 0], axis=3), axis=4)
points_X_for_summation = self._expand_torch(points_X_for_summation, points[:, :, :, 0].shape + (1, N))
points_Y_for_summation = np.expand_dims(np.expand_dims(points[:, :, :, 1], axis=3), axis=4)
points_Y_for_summation = self._expand_torch(points_Y_for_summation, points[:, :, :, 0].shape + (1, N))
if points_b == 1:
delta_X = points_X_for_summation - P_X
delta_Y = points_Y_for_summation - P_Y
else:
delta_X = points_X_for_summation - self._expand_torch(P_X, points_X_for_summation.shape)
delta_Y = points_Y_for_summation - self._expand_torch(P_Y, points_Y_for_summation.shape)
dist_squared = np.power(delta_X, 2) + np.power(delta_Y, 2)
dist_squared[dist_squared == 0] = 1
U = np.multiply(dist_squared, np.log(dist_squared))
points_X_batch = np.expand_dims(points[:,:,:,0], axis=3)
points_Y_batch = np.expand_dims(points[:,:,:,1], axis=3)
if points_b == 1:
points_X_batch = self._expand_torch(points_X_batch, (batch_size, ) + points_X_batch.shape[1:])
points_Y_batch = self._expand_torch(points_Y_batch, (batch_size, ) + points_Y_batch.shape[1:])
points_X_prime = A_X[:,:,:,:,0]+ \
np.multiply(A_X[:,:,:,:,1], points_X_batch) + \
np.multiply(A_X[:,:,:,:,2], points_Y_batch) + \
np.sum(np.multiply(W_X, self._expand_torch(U, W_X.shape)), 4)
points_Y_prime = A_Y[:,:,:,:,0]+ \
np.multiply(A_Y[:,:,:,:,1], points_X_batch) + \
np.multiply(A_Y[:,:,:,:,2], points_Y_batch) + \
np.sum(np.multiply(W_Y, self._expand_torch(U, W_Y.shape)), 4)
return np.concatenate((points_X_prime, points_Y_prime), 3)
def _generate_grid(self, theta):
grid_X, grid_Y, N, P_X, P_Y = self._prepare_to_transform()
warped_grid = self._apply_transformation(theta, np.concatenate((grid_X, grid_Y), axis=3), N, P_X, P_Y)
return warped_grid
def _bilinear_sampler(self, img, grid):
x, y = grid[:,:,:,0], grid[:,:,:,1]
H = img.shape[2]
W = img.shape[3]
max_y = H - 1
max_x = W - 1
# rescale x and y to [0, W-1/H-1]
x = 0.5 * (x + 1.0) * (max_x - 1)
y = 0.5 * (y + 1.0) * (max_y - 1)
# grab 4 nearest corner points for each (x_i, y_i)
x0 = np.floor(x).astype(int)
x1 = x0 + 1
y0 = np.floor(y).astype(int)
y1 = y0 + 1
# calculate deltas
wa = (x1 - x) * (y1 - y)
wb = (x1 - x) * (y - y0)
wc = (x - x0) * (y1 - y)
wd = (x - x0) * (y - y0)
# clip to range [0, H-1/W-1] to not violate img boundaries
x0 = np.clip(x0, 0, max_x)
x1 = np.clip(x1, 0, max_x)
y0 = np.clip(y0, 0, max_y)
y1 = np.clip(y1, 0, max_y)
# get pixel value at corner coords
img = img.reshape(-1, H, W)
Ia = img[:, y0, x0].swapaxes(0, 1)
Ib = img[:, y1, x0].swapaxes(0, 1)
Ic = img[:, y0, x1].swapaxes(0, 1)
Id = img[:, y1, x1].swapaxes(0, 1)
wa = np.expand_dims(wa, axis=0)
wb = np.expand_dims(wb, axis=0)
wc = np.expand_dims(wc, axis=0)
wd = np.expand_dims(wd, axis=0)
# compute output
out = wa*Ia + wb*Ib + wc*Ic + wd*Id
return out
class CorrelationLayer(object):
def __init__(self, params, blobs):
super(CorrelationLayer, self).__init__()
def getMemoryShapes(self, inputs):
fetureAShape = inputs[0]
b, c, h, w = fetureAShape
return [[b, h * w, h, w]]
def forward(self, inputs):
feature_A, feature_B = inputs
b, c, h, w = feature_A.shape
feature_A = feature_A.transpose(0, 1, 3, 2)
feature_A = np.reshape(feature_A, (b, c, h * w))
feature_B = np.reshape(feature_B, (b, c, h * w))
feature_B = feature_B.transpose(0, 2, 1)
feature_mul = feature_B @ feature_A
feature_mul= np.reshape(feature_mul, (b, h, w, h * w))
feature_mul = feature_mul.transpose(0, 1, 3, 2)
correlation_tensor = feature_mul.transpose(0, 2, 1, 3)
correlation_tensor = np.ascontiguousarray(correlation_tensor)
return [correlation_tensor]
if __name__ == "__main__":
if not os.path.isfile(args.gmm_model):
raise OSError("GMM model not exist")
if not os.path.isfile(args.tom_model):
raise OSError("TOM model not exist")
if not os.path.isfile(args.segmentation_model):
raise OSError("Segmentation model not exist")
if not os.path.isfile(findFile(args.openpose_proto)):
raise OSError("OpenPose proto not exist")
if not os.path.isfile(findFile(args.openpose_model)):
raise OSError("OpenPose model not exist")
person_img = cv.imread(args.input_image)
ratio = 256 / 192
inp_h, inp_w, _ = person_img.shape
current_ratio = inp_h / inp_w
if current_ratio > ratio:
center_h = inp_h // 2
out_h = inp_w * ratio
start = int(center_h - out_h // 2)
end = int(center_h + out_h // 2)
person_img = person_img[start:end, ...]
else:
center_w = inp_w // 2
out_w = inp_h / ratio
start = int(center_w - out_w // 2)
end = int(center_w + out_w // 2)
person_img = person_img[:, start:end, :]
cloth_img = cv.imread(args.input_cloth)
pose = get_pose_map(person_img, findFile(args.openpose_proto),
findFile(args.openpose_model), args.backend, args.target)
segm_image = parse_human(person_img, args.segmentation_model)
segm_image = cv.resize(segm_image, (192, 256), cv.INTER_LINEAR)
cv.dnn_registerLayer('Correlation', CorrelationLayer)
model = CpVton(args.gmm_model, args.tom_model, args.backend, args.target)
agnostic = model.prepare_agnostic(segm_image, person_img, pose)
warped_cloth = model.get_warped_cloth(cloth_img, agnostic)
output = model.get_tryon(agnostic, warped_cloth)
cv.dnn_unregisterLayer('Correlation')
winName = 'Virtual Try-On'
cv.namedWindow(winName, cv.WINDOW_AUTOSIZE)
cv.imshow(winName, output)
cv.waitKey()

191
samples/python/drawing.py Normal file
View File

@ -0,0 +1,191 @@
#!/usr/bin/env python
'''
This program demonstrates OpenCV drawing and text output functions by drawing different shapes and text strings
Usage :
python3 drawing.py
Press any button to exit
'''
# Python 2/3 compatibility
from __future__ import print_function
import numpy as np
import cv2 as cv
# Drawing Lines
def lines():
for i in range(NUMBER*2):
pt1, pt2 = [], []
pt1.append(np.random.randint(x1, x2))
pt1.append(np.random.randint(y1, y2))
pt2.append(np.random.randint(x1, x2))
pt2.append(np.random.randint(y1, y2))
color = "%06x" % np.random.randint(0, 0xFFFFFF)
color = tuple(int(color[i:i+2], 16) for i in (0, 2 ,4))
arrowed = np.random.randint(0, 6)
if (arrowed<3):
cv.line(image, tuple(pt1), tuple(pt2), color, np.random.randint(1, 10), lineType)
else:
cv.arrowedLine(image, tuple(pt1), tuple(pt2), color, np.random.randint(1, 10), lineType)
cv.imshow(wndname, image)
if cv.waitKey(DELAY)>=0:
return
# Drawing Rectangle
def rectangle():
for i in range(NUMBER*2):
pt1, pt2 = [], []
pt1.append(np.random.randint(x1, x2))
pt1.append(np.random.randint(y1, y2))
pt2.append(np.random.randint(x1, x2))
pt2.append(np.random.randint(y1, y2))
color = "%06x" % np.random.randint(0, 0xFFFFFF)
color = tuple(int(color[i:i+2], 16) for i in (0, 2 ,4))
thickness = np.random.randint(-3, 10)
marker = np.random.randint(0, 10)
marker_size = np.random.randint(30, 80)
if (marker > 5):
cv.rectangle(image, tuple(pt1), tuple(pt2), color, max(thickness, -1), lineType)
else:
cv.drawMarker(image, tuple(pt1), color, marker, marker_size)
cv.imshow(wndname, image)
if cv.waitKey(DELAY)>=0:
return
# Drawing ellipse
def ellipse():
for i in range(NUMBER*2):
center = []
center.append(np.random.randint(x1, x2))
center.append(np.random.randint(x1, x2))
axes = []
axes.append(np.random.randint(0, 200))
axes.append(np.random.randint(0, 200))
angle = np.random.randint(0, 180)
color = "%06x" % np.random.randint(0, 0xFFFFFF)
color = tuple(int(color[i:i+2], 16) for i in (0, 2 ,4))
thickness = np.random.randint(-1, 9)
cv.ellipse(image, tuple(center), tuple(axes), angle, angle-100, angle + 200, color, thickness, lineType)
cv.imshow(wndname, image)
if cv.waitKey(DELAY)>=0:
return
# Drawing Polygonal Curves
def polygonal():
for i in range(NUMBER):
pt = [(0, 0)]*6
pt = np.resize(pt, (2, 3, 2))
pt[0][0][0] = np.random.randint(x1, x2)
pt[0][0][1] = np.random.randint(y1, y2)
pt[0][1][0] = np.random.randint(x1, x2)
pt[0][1][1] = np.random.randint(y1, y2)
pt[0][2][0] = np.random.randint(x1, x2)
pt[0][2][1] = np.random.randint(y1, y2)
pt[1][0][0] = np.random.randint(x1, x2)
pt[1][0][1] = np.random.randint(y1, y2)
pt[1][1][0] = np.random.randint(x1, x2)
pt[1][1][1] = np.random.randint(y1, y2)
pt[1][2][0] = np.random.randint(x1, x2)
pt[1][2][1] = np.random.randint(y1, y2)
color = "%06x" % np.random.randint(0, 0xFFFFFF)
color = tuple(int(color[i:i+2], 16) for i in (0, 2 ,4))
alist = []
for k in pt[0]:
alist.append(k)
for k in pt[1]:
alist.append(k)
ppt = np.array(alist)
cv.polylines(image, [ppt], True, color, thickness = np.random.randint(1, 10), lineType = lineType)
cv.imshow(wndname, image)
if cv.waitKey(DELAY) >= 0:
return
# fills an area bounded by several polygonal contours
def fill():
for i in range(NUMBER):
pt = [(0, 0)]*6
pt = np.resize(pt, (2, 3, 2))
pt[0][0][0] = np.random.randint(x1, x2)
pt[0][0][1] = np.random.randint(y1, y2)
pt[0][1][0] = np.random.randint(x1, x2)
pt[0][1][1] = np.random.randint(y1, y2)
pt[0][2][0] = np.random.randint(x1, x2)
pt[0][2][1] = np.random.randint(y1, y2)
pt[1][0][0] = np.random.randint(x1, x2)
pt[1][0][1] = np.random.randint(y1, y2)
pt[1][1][0] = np.random.randint(x1, x2)
pt[1][1][1] = np.random.randint(y1, y2)
pt[1][2][0] = np.random.randint(x1, x2)
pt[1][2][1] = np.random.randint(y1, y2)
color = "%06x" % np.random.randint(0, 0xFFFFFF)
color = tuple(int(color[i:i+2], 16) for i in (0, 2 ,4))
alist = []
for k in pt[0]:
alist.append(k)
for k in pt[1]:
alist.append(k)
ppt = np.array(alist)
cv.fillPoly(image, [ppt], color, lineType)
cv.imshow(wndname, image)
if cv.waitKey(DELAY) >= 0:
return
# Drawing Circles
def circles():
for i in range(NUMBER):
center = []
center.append(np.random.randint(x1, x2))
center.append(np.random.randint(x1, x2))
color = "%06x" % np.random.randint(0, 0xFFFFFF)
color = tuple(int(color[i:i+2], 16) for i in (0, 2 ,4))
cv.circle(image, tuple(center), np.random.randint(0, 300), color, np.random.randint(-1, 9), lineType)
cv.imshow(wndname, image)
if cv.waitKey(DELAY) >= 0:
return
# Draws a text string
def string():
for i in range(NUMBER):
org = []
org.append(np.random.randint(x1, x2))
org.append(np.random.randint(x1, x2))
color = "%06x" % np.random.randint(0, 0xFFFFFF)
color = tuple(int(color[i:i+2], 16) for i in (0, 2 ,4))
cv.putText(image, "Testing text rendering", tuple(org), np.random.randint(0, 8), np.random.randint(0, 100)*0.05+0.1, color, np.random.randint(1, 10), lineType)
cv.imshow(wndname, image)
if cv.waitKey(DELAY) >= 0:
return
def string1():
textsize = cv.getTextSize("OpenCV forever!", cv.FONT_HERSHEY_COMPLEX, 3, 5)
org = (int((width - textsize[0][0])/2), int((height - textsize[0][1])/2))
for i in range(0, 255, 2):
image2 = np.array(image) - i
cv.putText(image2, "OpenCV forever!", org, cv.FONT_HERSHEY_COMPLEX, 3, (i, i, 255), 5, lineType)
cv.imshow(wndname, image2)
if cv.waitKey(DELAY) >= 0:
return
if __name__ == '__main__':
print(__doc__)
wndname = "Drawing Demo"
NUMBER = 100
DELAY = 5
width, height = 1000, 700
lineType = cv.LINE_AA # change it to LINE_8 to see non-antialiased graphics
x1, x2, y1, y2 = -width/2, width*3/2, -height/2, height*3/2
image = np.zeros((height, width, 3), dtype = np.uint8)
cv.imshow(wndname, image)
cv.waitKey(DELAY)
lines()
rectangle()
ellipse()
polygonal()
fill()
circles()
string()
string1()
cv.waitKey(0)
cv.destroyAllWindows()

16
samples/python/grabcut.py
View File

@ -11,10 +11,10 @@ USAGE:
README FIRST:
Two windows will show up, one for input and one for output.
At first, in input window, draw a rectangle around the object using
mouse right button. Then press 'n' to segment the object (once or a few times)
At first, in input window, draw a rectangle around the object using the
right mouse button. Then press 'n' to segment the object (once or a few times)
For any finer touch-ups, you can press any of the keys below and draw lines on
the areas you want. Then again press 'n' for updating the output.
the areas you want. Then again press 'n' to update the output.
Key '0' - To select areas of sure background
Key '1' - To select areas of sure foreground
@ -44,8 +44,8 @@ class App():
DRAW_BG = {'color' : BLACK, 'val' : 0}
DRAW_FG = {'color' : WHITE, 'val' : 1}
DRAW_PR_FG = {'color' : GREEN, 'val' : 3}
DRAW_PR_BG = {'color' : RED, 'val' : 2}
DRAW_PR_FG = {'color' : GREEN, 'val' : 3}
# setting up flags
rect = (0,0,1,1)
@ -160,14 +160,12 @@ class App():
print(""" For finer touchups, mark foreground and background after pressing keys 0-3
and again press 'n' \n""")
try:
bgdmodel = np.zeros((1, 65), np.float64)
fgdmodel = np.zeros((1, 65), np.float64)
if (self.rect_or_mask == 0): # grabcut with rect
bgdmodel = np.zeros((1, 65), np.float64)
fgdmodel = np.zeros((1, 65), np.float64)
cv.grabCut(self.img2, self.mask, self.rect, bgdmodel, fgdmodel, 1, cv.GC_INIT_WITH_RECT)
self.rect_or_mask = 1
elif self.rect_or_mask == 1: # grabcut with mask
bgdmodel = np.zeros((1, 65), np.float64)
fgdmodel = np.zeros((1, 65), np.float64)
elif (self.rect_or_mask == 1): # grabcut with mask
cv.grabCut(self.img2, self.mask, self.rect, bgdmodel, fgdmodel, 1, cv.GC_INIT_WITH_MASK)
except:
import traceback

69
samples/python/laplace.py Normal file
View File

@ -0,0 +1,69 @@
#!/usr/bin/env python
'''
This program demonstrates Laplace point/edge detection using
OpenCV function Laplacian()
It captures from the camera of your choice: 0, 1, ... default 0
Usage:
python laplace.py <ddepth> <smoothType> <sigma>
If no arguments given default arguments will be used.
Keyboard Shortcuts:
Press space bar to exit the program.
'''
# Python 2/3 compatibility
from __future__ import print_function
import numpy as np
import cv2 as cv
import sys
def main():
# Declare the variables we are going to use
ddepth = cv.CV_16S
smoothType = "MedianBlur"
sigma = 3
if len(sys.argv)==4:
ddepth = sys.argv[1]
smoothType = sys.argv[2]
sigma = sys.argv[3]
# Taking input from the camera
cap=cv.VideoCapture(0)
# Create Window and Trackbar
cv.namedWindow("Laplace of Image", cv.WINDOW_AUTOSIZE)
cv.createTrackbar("Kernel Size Bar", "Laplace of Image", sigma, 15, lambda x:x)
# Printing frame width, height and FPS
print("=="*40)
print("Frame Width: ", cap.get(cv.CAP_PROP_FRAME_WIDTH), "Frame Height: ", cap.get(cv.CAP_PROP_FRAME_HEIGHT), "FPS: ", cap.get(cv.CAP_PROP_FPS))
while True:
# Reading input from the camera
ret, frame = cap.read()
if ret == False:
print("Can't open camera/video stream")
break
# Taking input/position from the trackbar
sigma = cv.getTrackbarPos("Kernel Size Bar", "Laplace of Image")
# Setting kernel size
ksize = (sigma*5)|1
# Removing noise by blurring with a filter
if smoothType == "GAUSSIAN":
smoothed = cv.GaussianBlur(frame, (ksize, ksize), sigma, sigma)
if smoothType == "BLUR":
smoothed = cv.blur(frame, (ksize, ksize))
if smoothType == "MedianBlur":
smoothed = cv.medianBlur(frame, ksize)
# Apply Laplace function
laplace = cv.Laplacian(smoothed, ddepth, 5)
# Converting back to uint8
result = cv.convertScaleAbs(laplace, (sigma+1)*0.25)
# Display Output
cv.imshow("Laplace of Image", result)
k = cv.waitKey(30)
if k == 27:
return
if __name__ == "__main__":
print(__doc__)
main()
cv.destroyAllWindows()