void procOCL_OCV(int tex, int w, int h)
{
int64_t t = getTimeMs();
cl::ImageGL imgIn (theContext, CL_MEM_READ_ONLY, GL_TEXTURE_2D, 0, tex);
std::vector < cl::Memory > images(1, imgIn);
theQueue.enqueueAcquireGLObjects(&images);
theQueue.finish();
cv::UMat uIn, uOut, uTmp;
cv::ocl::convertFromImage(imgIn(), uIn);
LOGD("loading texture data to OpenCV UMat costs %d ms", getTimeInterval(t));
theQueue.enqueueReleaseGLObjects(&images);
t = getTimeMs();
//cv::blur(uIn, uOut, cv::Size(5, 5));
cv::Laplacian(uIn, uTmp, CV_8U);
cv:multiply(uTmp, 10, uOut);
cv::ocl::finish();
LOGD("OpenCV processing costs %d ms", getTimeInterval(t));
t = getTimeMs();
cl::ImageGL imgOut(theContext, CL_MEM_WRITE_ONLY, GL_TEXTURE_2D, 0, tex);
images.clear();
images.push_back(imgOut);
theQueue.enqueueAcquireGLObjects(&images);
cl_mem clBuffer = (cl_mem)uOut.handle(cv::ACCESS_READ);
cl_command_queue q = (cl_command_queue)cv::ocl::Queue::getDefault().ptr();
size_t offset = 0;
size_t origin[3] = { 0, 0, 0 };
size_t region[3] = { w, h, 1 };
CV_Assert(clEnqueueCopyBufferToImage (q, clBuffer, imgOut(), offset, origin, region, 0, NULL, NULL) == CL_SUCCESS);
theQueue.enqueueReleaseGLObjects(&images);
cv::ocl::finish();
LOGD("uploading results to texture costs %d ms", getTimeInterval(t));
}
void procOCL_I2I(int texIn, int texOut, int w, int h)
{
if(!haveOpenCL) return;
LOGD("procOCL_I2I(%d, %d, %d, %d)", texIn, texOut, w, h);
cl::ImageGL imgIn (theContext, CL_MEM_READ_ONLY, GL_TEXTURE_2D, 0, texIn);
cl::ImageGL imgOut(theContext, CL_MEM_WRITE_ONLY, GL_TEXTURE_2D, 0, texOut);
std::vector < cl::Memory > images;
images.push_back(imgIn);
images.push_back(imgOut);
int64_t t = getTimeMs();
theQueue.enqueueAcquireGLObjects(&images);
theQueue.finish();
LOGD("enqueueAcquireGLObjects() costs %d ms", getTimeInterval(t));
t = getTimeMs();
cl::Kernel Laplacian(theProgI2I, "Laplacian"); //TODO: may be done once
Laplacian.setArg(0, imgIn);
Laplacian.setArg(1, imgOut);
theQueue.finish();
LOGD("Kernel() costs %d ms", getTimeInterval(t));
t = getTimeMs();
theQueue.enqueueNDRangeKernel(Laplacian, cl::NullRange, cl::NDRange(w, h), cl::NullRange);
theQueue.finish();
LOGD("enqueueNDRangeKernel() costs %d ms", getTimeInterval(t));
t = getTimeMs();
theQueue.enqueueReleaseGLObjects(&images);
theQueue.finish();
LOGD("enqueueReleaseGLObjects() costs %d ms", getTimeInterval(t));
}
/**
* @brief Execute
* @param nameIn
* @param nameOut
* @param size
* @return
*/
static Image *Execute(std::string nameIn, std::string nameOut, int size)
{
Image imgIn(nameIn);
Image *imgOut = Execute(&imgIn, NULL, size);
imgOut->Write(nameOut);
return imgOut;
}
/**
* @brief Execute
* @param fileInput
* @param fileOutput
* @return
*/
static Image *Execute(std::string fileInput, std::string fileOutput)
{
Image imgIn(fileInput);
Image *out = FilterNSWE::Execute(&imgIn, NULL);
out->Write(fileOutput);
return out;
}
/**
* @brief Execute
* @param fileInput
* @param fileOutput
* @param min
* @param max
* @return
*/
static Image *Execute(std::string fileInput, std::string fileOutput,
Vec2i min, Vec2i max)
{
Image imgIn(fileInput);
Image *out = FilterCrop::Execute(&imgIn, NULL, min, max);
out->Write(fileOutput);
return out;
}
/**
* @brief Execute
* @param nameFileIn
* @param nameFileOut
* @param threshold_nuked
* @return
*/
static Image* Execute(std::string nameFileIn, std::string nameFileOut, float threshold_nuked = 1e4)
{
Image imgIn(nameFileIn);
Image *imgOut = Execute(&imgIn, NULL, threshold_nuked);
imgOut->Write(nameFileOut);
return imgOut;
}
void SCropVideoQuad::cropFrame(::fwCore::HiResClock::HiResClockType timestamp)
{
::arData::FrameTL::csptr frameTL = this->getInput< ::arData::FrameTL >(s_FRAMETL_INPUT);
::fwCore::HiResClock::HiResClockType newerTimestamp = frameTL->getNewerTimestamp();
const CSPTR(::arData::FrameTL::BufferType) buffer = frameTL->getClosestBuffer(newerTimestamp);
::arData::FrameTL::sptr frameTL1 = this->getInOut< ::arData::FrameTL >(s_FRAMETL1_INOUT);
::arData::FrameTL::sptr frameTL2 = this->getInOut< ::arData::FrameTL >(s_FRAMETL2_INOUT);
::arData::FrameTL::sptr frameTL3 = this->getInOut< ::arData::FrameTL >(s_FRAMETL3_INOUT);
::arData::FrameTL::sptr frameTL4 = this->getInOut< ::arData::FrameTL >(s_FRAMETL4_INOUT);
const size_t width = frameTL->getWidth();
const size_t height = frameTL->getHeight();
const size_t widthTL1 = frameTL1->getWidth();
const size_t heightTL1 = frameTL1->getHeight();
const size_t widthCropped = width / 2;
const size_t heightCropped = height / 2;
if (widthCropped != widthTL1 || heightCropped != heightTL1)
{
const size_t numberOfComponents = frameTL->getNumberOfComponents();
const ::fwTools::Type type = frameTL->getType();
frameTL1->clearTimeline();
frameTL1->initPoolSize(widthCropped, heightCropped, type, numberOfComponents);
frameTL2->clearTimeline();
frameTL2->initPoolSize(widthCropped, heightCropped, type, numberOfComponents);
frameTL3->clearTimeline();
frameTL3->initPoolSize(widthCropped, heightCropped, type, numberOfComponents);
frameTL4->clearTimeline();
frameTL4->initPoolSize(widthCropped, heightCropped, type, numberOfComponents);
}
if(buffer)
{
const std::uint8_t* frameBuff = &buffer->getElement(0);
// The cv image that will be processed
::cv::Mat imgIn(height, width, CV_8UC4, (void*)frameBuff, ::cv::Mat::AUTO_STEP);
const ::cv::Rect roi1(0, 0, widthCropped, heightCropped);
const ::cv::Rect roi2(widthCropped, 0, widthCropped, heightCropped);
const ::cv::Rect roi3(0, heightCropped, widthCropped, heightCropped);
const ::cv::Rect roi4(widthCropped, heightCropped, widthCropped, heightCropped);
this->pushFrameInTimeline(imgIn, roi1, frameTL1, newerTimestamp);
this->pushFrameInTimeline(imgIn, roi2, frameTL2, newerTimestamp);
this->pushFrameInTimeline(imgIn, roi3, frameTL3, newerTimestamp);
this->pushFrameInTimeline(imgIn, roi4, frameTL4, newerTimestamp);
}
}
/**
* @brief execute
* @param nameIn
* @param nameOut
* @param sigma_s
* @param sigma_r
* @param testing
* @return
*/
static ImageGL *execute(std::string nameIn, std::string nameOut,
float sigma_s, float sigma_r, int testing)
{
GLuint testTQ = glBeginTimeQuery();
glEndTimeQuery(testTQ);
ImageGL imgIn(nameIn);
imgIn.generateTextureGL(GL_TEXTURE_2D, GL_FLOAT, false);
FilterGLBilateral2DSP filter(sigma_s, sigma_r);
ImageGL *imgRet = new ImageGL(1, imgIn.width, imgIn.height, 3, IMG_GPU, GL_TEXTURE_2D);
GLuint testTQ1;
if(testing > 1) {
filter.Process(SingleGL(&imgIn), imgRet);
testTQ1 = glBeginTimeQuery();
for(int i = 0; i < testing; i++) {
filter.Process(SingleGL(&imgIn), imgRet);
}
} else {
testTQ1 = glBeginTimeQuery();
filter.Process(SingleGL(&imgIn), imgRet);
}
GLuint64EXT timeVal = glEndTimeQuery(testTQ1);
double ms = double(timeVal) / (double(testing) * 1000000.0);
printf("Separate Bilateral Filter on GPU time: %f ms\n", ms);
std::string sign = genBilString("S", sigma_s, sigma_r);
std::string nameTime = FileLister::getFileNumber(sign, "txt");
FILE *file = fopen(nameTime.c_str(), "w");
if(file != NULL) {
fprintf(file, "%f", ms);
fclose(file);
}
ImageGL *imgWrite = new ImageGL(1, imgIn.width, imgIn.height, 4, IMG_CPU, GL_TEXTURE_2D);
imgWrite->readFromFBO(filter.getFbo());
imgWrite->Write(nameOut);
return imgRet;
}
/**
* @brief Test
*/
static void Test()
{
Image imgIn(1, 512, 512, 3);
imgIn = 1.0f;
FilterChannel filter(0);
Image *outR = filter.Process(Single(&imgIn), NULL);
filter.setChannel(1);
Image *outG = filter.Process(Single(&imgIn), NULL);
filter.setChannel(2);
Image *outB = filter.Process(Single(&imgIn), NULL);
outR->Write("channel_R.pfm");
outG->Write("channel_G.pfm");
outB->Write("channel_B.pfm");
}
/**
* @brief ExecuteTest
* @param nameIn
* @param nameOut
* @return
*/
static Image *ExecuteTest(std::string nameIn, std::string nameOut)
{
Image imgIn(nameIn);
FilterChannel filter(0);
Image *outR = filter.Process(Single(&imgIn), NULL);
filter.setChannel(1);
Image *outG = filter.Process(Single(&imgIn), NULL);
filter.setChannel(2);
Image *outB = filter.Process(Single(&imgIn), NULL);
ImageVec src;
src.push_back(outR);
src.push_back(outG);
src.push_back(outB);
FilterCombine filterC;
Image *ret = filterC.Process(src, NULL);
ret->Write(nameOut);
return ret;
}
