bool YuvToAnyAutoTest(int width, int height, bool is420, View::Format dstType, const Func & f1, const Func & f2)
{
bool result = true;
std::cout << "Test " << f1.description << " & " << f2.description << " [" << width << ", " << height << "]." << std::endl;
const int uvWidth = is420 ? width/2 : width;
const int uvHeight = is420 ? height/2 : height;
View y(width, height, View::Gray8, NULL, TEST_ALIGN(width));
FillRandom(y);
View u(uvWidth, uvHeight, View::Gray8, NULL, TEST_ALIGN(uvWidth));
FillRandom(u);
View v(uvWidth, uvHeight, View::Gray8, NULL, TEST_ALIGN(uvWidth));
FillRandom(v);
View dst1(width, height, dstType, NULL, TEST_ALIGN(width));
View dst2(width, height, dstType, NULL, TEST_ALIGN(width));
TEST_EXECUTE_AT_LEAST_MIN_TIME(f1.Call(y, u, v, dst1));
TEST_EXECUTE_AT_LEAST_MIN_TIME(f2.Call(y, u, v, dst2));
result = result && Compare(dst1, dst2, 0, true, 64);
return result;
}
int main(int argc, char *argv[])
{
//if (argc != 3) return 1;
IoChannel src1(argv[1], "r");
IoChannel dst1(argv[2], "w");
IoChannel src2(argv[1], "r");
IoChannel dst2(argv[3], "w");
CopyChannel cc1(src1, dst1);
CopyChannel cc2(src2, dst2, 2000);
GMainLoop *loop = g_main_loop_new(NULL, FALSE);
g_main_loop_run(loop);
return 0;
}
TEST(algorithmTest, transform_image)
{
const int width = 10;
const int height = 10;
cv::Mat_<int> src1(height, width);
for(int y=0; y<src1.rows; ++y) {
for(int x=0; x<src1.cols; ++x) {
src1(y, x) = y*src1.cols + x;
}
}
cv::Mat_<int> src2(height, width);
for(int y=0; y<src2.rows; ++y) {
for(int x=0; x<src2.cols; ++x) {
src2(y, x) = x*src2.cols + y;
}
}
cv::Mat_<int> src3(height, width);
for(int y=0; y<src3.rows; ++y) {
for(int x=0; x<src3.cols; ++x) {
src3(y, x) = x*y;
}
}
cv::Mat_<int> dst1(height, width);
cv::Mat_<int> dst2(height, width);
cvutil::transform_image(src1, dst1, [](int x){ return x*2; });
for(int y=0; y<dst2.rows; ++y) {
for(int x=0; x<dst2.cols; ++x) {
dst2(y, x) = src1(y, x)*2;
}
}
ASSERT_TRUE(std::equal(dst2.begin(), dst2.end(), dst2.begin()));
cvutil::transform_image(src1, src2, dst1,
[](int x, int y){ return x*2 + y*2; });
for(int y=0; y<dst2.rows; ++y) {
for(int x=0; x<dst2.cols; ++x) {
dst2(y, x) = src1(y, x)*2 + src2(y, x)*2;
}
}
ASSERT_TRUE(std::equal(dst2.begin(), dst2.end(), dst2.begin()));
cvutil::transform_image(src1, src2, src3, dst1,
[](int x, int y, int z){ return x + y + z; });
for(int y=0; y<dst2.rows; ++y) {
for(int x=0; x<dst2.cols; ++x) {
dst2(y, x) = src1(y, x) + src2(y, x) + src3(y, x);
}
}
ASSERT_TRUE(std::equal(dst2.begin(), dst2.end(), dst2.begin()));
}
bool YuvToAnyDataTest(bool create, int width, int height, bool is420, View::Format dstType, const Func & f)
{
bool result = true;
Data data(f.description);
std::cout << (create ? "Create" : "Verify") << " test " << f.description << " [" << width << ", " << height << "]." << std::endl;
const int uvWidth = is420 ? width/2 : width;
const int uvHeight = is420 ? height/2 : height;
View y(width, height, View::Gray8, NULL, TEST_ALIGN(width));
View u(uvWidth, uvHeight, View::Gray8, NULL, TEST_ALIGN(uvWidth));
View v(uvWidth, uvHeight, View::Gray8, NULL, TEST_ALIGN(uvWidth));
View dst1(width, height, dstType, NULL, TEST_ALIGN(width));
View dst2(width, height, dstType, NULL, TEST_ALIGN(width));
if(create)
{
FillRandom(y);
FillRandom(u);
FillRandom(v);
TEST_SAVE(y);
TEST_SAVE(u);
TEST_SAVE(v);
f.Call(y, u, v, dst1);
TEST_SAVE(dst1);
}
else
{
TEST_LOAD(y);
TEST_LOAD(u);
TEST_LOAD(v);
TEST_LOAD(dst1);
f.Call(y, u, v, dst2);
TEST_SAVE(dst2);
result = result && Compare(dst1, dst2, 0, true, 64);
}
return result;
}
void copy_channels_node_t::do_process( const render::context_t& context)
{
Imath::Box2i src1_area( ImathExt::intersect( input_as<image_node_t>(0)->defined(), defined()));
image::const_image_view_t src1( input_as<image_node_t>(0)->const_subimage_view( src1_area));
Imath::Box2i src2_area( ImathExt::intersect( input_as<image_node_t>(1)->defined(), defined()));
image::const_image_view_t src2( input_as<image_node_t>(1)->const_subimage_view( src2_area));
image::image_view_t dst1( subimage_view( src1_area));
image::image_view_t dst2( subimage_view( src2_area));
int ch = get_value<int>( param( "red"));
if( ch == copy_source)
copy_channel( src1, 0, dst1, 0);
else
copy_channel( src2, ch-1, dst2, 0);
ch = get_value<int>( param( "green"));
if( ch == copy_source)
copy_channel( src1, 1, dst1, 1);
else
copy_channel( src2, ch-1, dst2, 1);
ch = get_value<int>( param( "blue"));
if( ch == copy_source)
copy_channel( src1, 2, dst1, 2);
else
copy_channel( src2, ch-1, dst2, 2);
ch = get_value<int>( param( "alpha"));
switch( ch)
{
case copy_source:
copy_channel( src1, 3, dst1, 3);
break;
case set_one:
case set_zero:
{
Imath::Box2i area( ImathExt::intersect( defined(), format()));
copy_channel( src1, ch-1, subimage_view( area), 3);
}
break;
default:
copy_channel( src2, ch-1, dst2, 3);
}
}
int test_layer_reverse()
{
// Blog: http://blog.csdn.net/fengbingchun/article/details/77160872
#ifdef __linux__
std::string image_name{ "test_data/images/lena.png" };
#else
std::string image_name{ "E:/GitCode/CUDA_Test/test_data/images/lena.png" };
#endif
cv::Mat matSrc = cv::imread(image_name);
CHECK(matSrc.data);
cv::cvtColor(matSrc, matSrc, CV_BGR2GRAY);
const int width{ 1511 }, height{ 1473 };
const auto length = width * height;
cv::resize(matSrc, matSrc, cv::Size(width, height));
cv::Mat matTmp1;
matSrc.convertTo(matTmp1, CV_32FC1);
float elapsed_time1{ 0.f }, elapsed_time2{ 0.f }; // milliseconds
const std::vector<int> vec{ 5, 7};
std::unique_ptr<float[]> dst1(new float[length]), dst2(new float[length]);
std::for_each(dst1.get(), dst1.get() + length, [](float& n) {n = 0.f; });
std::for_each(dst2.get(), dst2.get() + length, [](float& n) {n = 0.f; });
int ret = layer_reverse_cpu((float*)matTmp1.data, dst1.get(), length, vec, &elapsed_time1);
if (ret != 0) PRINT_ERROR_INFO(image_reverse_cpu);
ret = layer_reverse_gpu((float*)matTmp1.data, dst2.get(), length, vec, &elapsed_time2);
if (ret != 0) PRINT_ERROR_INFO(image_reverse_gpu);
compare_result(dst1.get(), dst2.get(), length);
cv::Mat matTmp2(height, width, CV_32FC1, dst2.get()), matDst;
matTmp2.convertTo(matDst, CV_8UC1);
#ifdef __linux__
save_image(matSrc, matDst, 400, 200, "test_data/images/image_reverse.png");
#else
save_image(matSrc, matDst, 400, 200, "E:/GitCode/CUDA_Test/test_data/images/image_reverse.png");
#endif
fprintf(stderr, "test layer reverse: cpu run time: %f ms, gpu run time: %f ms\n", elapsed_time1, elapsed_time2);
return 0;
}
bool AveragingBinarizationDataTest(bool create, int width, int height, SimdCompareType type, const Func2 & f)
{
bool result = true;
Data data(f.description);
TEST_LOG_SS(Info, (create ? "Create" : "Verify") << " test " << f.description << " [" << width << ", " << height << "].");
View src(width, height, View::Gray8, NULL, TEST_ALIGN(width));
View dst1(width, height, View::Gray8, NULL, TEST_ALIGN(width));
View dst2(width, height, View::Gray8, NULL, TEST_ALIGN(width));
const uint8_t value = 127;
const size_t neighborhood = 17;
const uint8_t threshold = 128;
const uint8_t positive = 7;
const uint8_t negative = 3;
if(create)
{
FillRandom(src);
TEST_SAVE(src);
f.Call(src, value, neighborhood, threshold, positive, negative, dst1, type);
TEST_SAVE(dst1);
}
else
{
TEST_LOAD(src);
TEST_LOAD(dst1);
f.Call(src, value, neighborhood, threshold, positive, negative, dst2, type);
TEST_SAVE(dst2);
result = result && Compare(dst1, dst2, 0, true, 64);
}
return result;
}
bool ContourMetricsMaskedDataTest(bool create, int width, int height, const FuncM & f)
{
bool result = true;
Data data(f.description);
TEST_LOG_SS(Info, (create ? "Create" : "Verify") << " test " << f.description << " [" << width << ", " << height << "].");
View src(width, height, View::Gray8, NULL, TEST_ALIGN(width));
View mask(width, height, View::Gray8, NULL, TEST_ALIGN(width));
View dst1(width, height, View::Int16, NULL, TEST_ALIGN(width));
View dst2(width, height, View::Int16, NULL, TEST_ALIGN(width));
if(create)
{
FillRandom(src);
FillRandom(mask);
TEST_SAVE(src);
TEST_SAVE(mask);
f.Call(src, mask, 128, dst1);
TEST_SAVE(dst1);
}
else
{
TEST_LOAD(src);
TEST_LOAD(mask);
TEST_LOAD(dst1);
f.Call(src, mask, 128, dst2);
TEST_SAVE(dst2);
result = result && Compare(dst1, dst2, 0, true, 32, 0);
}
return result;
}
bool AnyToAnyAutoTest(int width, int height, View::Format srcType, View::Format dstType, const Func & f1, const Func & f2)
{
bool result = true;
TEST_LOG_SS(Info, "Test " << f1.description << " & " << f2.description << " for size [" << width << "," << height << "].");
View src(width, height, srcType, NULL, TEST_ALIGN(width));
FillRandom(src);
View dst1(width, height, dstType, NULL, TEST_ALIGN(width));
View dst2(width, height, dstType, NULL, TEST_ALIGN(width));
TEST_EXECUTE_AT_LEAST_MIN_TIME(f1.Call(src, dst1));
TEST_EXECUTE_AT_LEAST_MIN_TIME(f2.Call(src, dst2));
result = result && Compare(dst1, dst2, 0, true, 64);
return result;
}
bool TexturePerformCompensationAutoTest(int width, int height, int shift, const Func4 & f1, const Func4 & f2)
{
bool result = true;
TEST_LOG_SS(Info, "Test " << f1.description << " & " << f2.description << " [" << width << ", " << height << "] <" << shift << ">.");
View src(width, height, View::Gray8, NULL, TEST_ALIGN(width));
FillRandom(src);
View dst1(width, height, View::Gray8, NULL, TEST_ALIGN(width));
View dst2(width, height, View::Gray8, NULL, TEST_ALIGN(width));
TEST_EXECUTE_AT_LEAST_MIN_TIME(f1.Call(src, shift, dst1));
TEST_EXECUTE_AT_LEAST_MIN_TIME(f2.Call(src, shift, dst2));
result = result && Compare(dst1, dst2, 0, true, 32, 0);
return result;
}
bool TexturePerformCompensationDataTest(bool create, int width, int height, const Func4 & f)
{
bool result = true;
Data data(f.description);
TEST_LOG_SS(Info, (create ? "Create" : "Verify") << " test " << f.description << " [" << width << ", " << height << "].");
View src(width, height, View::Gray8, NULL, TEST_ALIGN(width));
FillRandom(src);
View dst1(width, height, View::Gray8, NULL, TEST_ALIGN(width));
View dst2(width, height, View::Gray8, NULL, TEST_ALIGN(width));
const int shift = -7;
if(create)
{
FillRandom(src);
TEST_SAVE(src);
f.Call(src, shift, dst1);
TEST_SAVE(dst1);
}
else
{
TEST_LOAD(src);
TEST_LOAD(dst1);
f.Call(src, shift, dst2);
TEST_SAVE(dst2);
result = result && Compare(dst1, dst2, 0, true, 32, 0);
}
return result;
}
bool AnyToAnyDataTest(bool create, int width, int height, View::Format srcType, View::Format dstType, const Func & f)
{
bool result = true;
Data data(f.description);
TEST_LOG_SS(Info, (create ? "Create" : "Verify") << " test " << f.description << " [" << width << ", " << height << "].");
View src(width, height, srcType, NULL, TEST_ALIGN(width));
View dst1(width, height, dstType, NULL, TEST_ALIGN(width));
View dst2(width, height, dstType, NULL, TEST_ALIGN(width));
if(create)
{
FillRandom(src);
TEST_SAVE(src);
f.Call(src, dst1);
TEST_SAVE(dst1);
}
else
{
TEST_LOAD(src);
TEST_LOAD(dst1);
f.Call(src, dst2);
TEST_SAVE(dst2);
result = result && Compare(dst1, dst2, 0, true, 64, 0);
}
return result;
}
TEST(algorithmTest, transform_image_xy)
{
const int width = 10;
const int height = 10;
cv::Mat_<int> src1(height, width);
for(int y=0; y<src1.rows; ++y) {
for(int x=0; x<src1.cols; ++x) {
src1(y, x) = y * x;
}
}
cv::Mat_<int> src2(height, width);
for(int y=0; y<src2.rows; ++y) {
for(int x=0; x<src2.cols; ++x) {
src2(y, x) = 2*x + 3*y;
}
}
cv::Mat_<int> dst1(height, width);
cvutil::transform_image_xy(src1, dst1,
[](int p, int x, int y){ return p + x + 3*y; });
cv::Mat_<int> dst2(height, width);
for(int y=0; y<dst2.rows; ++y) {
for(int x=0; x<dst2.cols; ++x) {
dst2(y, x) = x*y + x + 3*y;
}
}
ASSERT_TRUE(std::equal(dst1.begin(), dst1.end(), dst2.begin()));
cvutil::transform_image_xy(src1, src2, dst1,
[](int p1, int p2, int x, int y){ return p1*p2 + x + 5*y; });
for(int y=0; y<dst2.rows; ++y) {
for(int x=0; x<dst2.cols; ++x) {
dst2(y, x) = x*y*(2*x + 3*y) + x + 5*y;
}
}
ASSERT_TRUE(std::equal(dst1.begin(), dst1.end(), dst2.begin()));
}
void dng_image::GetRepeat (dng_pixel_buffer &buffer,
const dng_rect &srcArea,
const dng_rect &dstArea) const
{
// If we already have the entire srcArea in the
// buffer, we can just repeat that.
if ((srcArea & buffer.fArea) == srcArea)
{
buffer.RepeatArea (srcArea,
dstArea);
}
// Else we first need to get the srcArea into the buffer area.
else
{
// Find repeating pattern size.
dng_point repeat = srcArea.Size ();
// Find pattern phase at top-left corner of destination area.
dng_point phase = dng_pixel_buffer::RepeatPhase (srcArea,
dstArea);
// Find new source area at top-left of dstArea.
dng_rect newArea = srcArea + (dstArea.TL () -
srcArea.TL ());
// Find quadrant split coordinates.
int32 splitV = newArea.t + repeat.v - phase.v;
int32 splitH = newArea.l + repeat.h - phase.h;
// Top-left quadrant.
dng_rect dst1 (dng_rect (newArea.t,
newArea.l,
splitV,
splitH) & dstArea);
if (dst1.NotEmpty ())
{
dng_pixel_buffer temp (buffer);
temp.fArea = dst1 + (srcArea.TL () -
dstArea.TL () +
dng_point (phase.v, phase.h));
temp.fData = buffer.DirtyPixel (dst1.t,
dst1.l,
buffer.fPlane);
DoGet (temp);
}
// Top-right quadrant.
dng_rect dst2 (dng_rect (newArea.t,
splitH,
splitV,
newArea.r) & dstArea);
if (dst2.NotEmpty ())
{
dng_pixel_buffer temp (buffer);
temp.fArea = dst2 + (srcArea.TL () -
dstArea.TL () +
dng_point (phase.v, -phase.h));
temp.fData = buffer.DirtyPixel (dst2.t,
dst2.l,
buffer.fPlane);
DoGet (temp);
}
// Bottom-left quadrant.
dng_rect dst3 (dng_rect (splitV,
newArea.l,
newArea.b,
splitH) & dstArea);
if (dst3.NotEmpty ())
{
dng_pixel_buffer temp (buffer);
temp.fArea = dst3 + (srcArea.TL () -
dstArea.TL () +
dng_point (-phase.v, phase.h));
temp.fData = buffer.DirtyPixel (dst3.t,
dst3.l,
buffer.fPlane);
DoGet (temp);
}
// Bottom-right quadrant.
dng_rect dst4 (dng_rect (splitV,
splitH,
newArea.b,
newArea.r) & dstArea);
if (dst4.NotEmpty ())
{
dng_pixel_buffer temp (buffer);
temp.fArea = dst4 + (srcArea.TL () -
dstArea.TL () +
dng_point (-phase.v, -phase.h));
temp.fData = buffer.DirtyPixel (dst4.t,
dst4.l,
buffer.fPlane);
DoGet (temp);
}
// Replicate this new source area.
buffer.RepeatArea (newArea,
dstArea);
}
}
ERcExitCode CResourceCompilerHelper::InvokeResourceCompiler(const char* szSrcFilePath, const char* szDstFilePath, const bool bUserDialog)
{
const char* szAdjustedFilename = szSrcFilePath;
#if defined(USE_GAMESTREAM)
char szFullSrcPathBuf[ICryPak::g_nMaxPath];
if (gEnv && gEnv->pGameStream)
{
szAdjustedFilename = gEnv->pCryPak->AdjustFileName(szSrcFilePath, szFullSrcPathBuf, ICryPak::FLAGS_RESOLVE_TO_CACHE);
}
#endif
ERcExitCode eRet = eRcExitCode_Pending;
// make command for execution
wchar_t szProjectDir[512];
GetCurrentDirectoryW(sizeof(szProjectDir) / sizeof(szProjectDir[0]), szProjectDir);
SettingsManagerHelpers::CFixedString<wchar_t, 512> wRemoteCmdLine;
SettingsManagerHelpers::CFixedString<wchar_t, 512> wDir;
CSettingsManagerTools smTools = CSettingsManagerTools();
const char* const szRcParentDir =
(smTools.Is64bitWindows() && (DirectoryExists(L"Bin64/rc") || !DirectoryExists(L"Bin32/rc")))
? "Bin64"
: "Bin32";
wchar_t szRegSettingsBuffer[1024];
smTools.GetEngineSettingsManager()->GetValueByRef("RC_Parameters", SettingsManagerHelpers::CWCharBuffer(szRegSettingsBuffer, sizeof(szRegSettingsBuffer)));
wRemoteCmdLine.appendAscii(szRcParentDir);
wRemoteCmdLine.appendAscii("/rc/");
wRemoteCmdLine.appendAscii(RC_EXECUTABLE);
wRemoteCmdLine.appendAscii(" \"");
wRemoteCmdLine.append(szProjectDir);
wRemoteCmdLine.appendAscii("\\");
wRemoteCmdLine.appendAscii(szAdjustedFilename);
wRemoteCmdLine.appendAscii("\" /userdialog=0 ");
wRemoteCmdLine.append(szRegSettingsBuffer);
// make it write to a filename of our choice
char szDstFilename[512];
char szDstPath[512];
RemovePath(szDstFilePath, szDstFilename, 512);
RemoveFilename(szDstFilePath, szDstPath, 512);
wRemoteCmdLine.appendAscii(" /overwritefilename=\"");
wRemoteCmdLine.appendAscii(szDstFilename);
wRemoteCmdLine.appendAscii("\"");
wRemoteCmdLine.appendAscii(" /targetroot=\"");
wRemoteCmdLine.append(szProjectDir);
wRemoteCmdLine.appendAscii("\\");
wRemoteCmdLine.appendAscii(szDstPath);
wRemoteCmdLine.appendAscii("\"");
wDir.append(szProjectDir);
wDir.appendAscii("\\");
wDir.appendAscii(szRcParentDir);
wDir.appendAscii("\\rc");
STARTUPINFOW si;
ZeroMemory(&si, sizeof(si));
si.cb = sizeof(si);
si.dwX = 100;
si.dwY = 100;
si.dwFlags = STARTF_USEPOSITION;
PROCESS_INFORMATION pi;
ZeroMemory(&pi, sizeof(pi));
#if defined(DEBUG) && !defined(NDEBUG) && defined(_RENDERER)
extern ILog *iLog;
char tmp1[512];
char tmp2[512];
SettingsManagerHelpers::CCharBuffer dst1(tmp1, 512); SettingsManagerHelpers::ConvertUtf16ToUtf8(wDir.c_str(), dst1);
SettingsManagerHelpers::CCharBuffer dst2(tmp2, 512); SettingsManagerHelpers::ConvertUtf16ToUtf8(wRemoteCmdLine.c_str(), dst2);
iLog->Log("Debug: RC: dir \"%s\", cmd \"%s\"\n", tmp1, tmp2);
#endif
if (!CreateProcessW(
NULL, // No module name (use command line).
const_cast<wchar_t*>(wRemoteCmdLine.c_str()), // Command line.
NULL, // Process handle not inheritable.
NULL, // Thread handle not inheritable.
FALSE, // Set handle inheritance to FALSE.
BELOW_NORMAL_PRIORITY_CLASS + (bUserDialog ? 0 : CREATE_NO_WINDOW), // creation flags.
NULL, // Use parent's environment block.
wDir.c_str(), // Set starting directory.
&si, // Pointer to STARTUPINFO structure.
&pi)) // Pointer to PROCESS_INFORMATION structure.
{
eRet = eRcExitCode_FatalError;
}
else
{
// Wait until child process exits.
WaitForSingleObject(pi.hProcess, INFINITE);
DWORD exitCode;
if (GetExitCodeProcess(pi.hProcess, &exitCode) == 0)
{
eRet = eRcExitCode_Error;
}
else
{
eRet = (ERcExitCode)exitCode;
}
}
// Close process and thread handles.
CloseHandle(pi.hProcess);
CloseHandle(pi.hThread);
return eRet;
}
void crossCorr( const Mat& img, const Mat& _templ, Mat& corr,
Size corrsize, int ctype,
Point anchor, double delta, int borderType )
{
const double blockScale = 4.5;
const int minBlockSize = 256;
std::vector<uchar> buf;
Mat templ = _templ;
int depth = img.depth(), cn = img.channels();
int tdepth = templ.depth(), tcn = templ.channels();
int cdepth = CV_MAT_DEPTH(ctype), ccn = CV_MAT_CN(ctype);
CV_Assert( img.dims <= 2 && templ.dims <= 2 && corr.dims <= 2 );
if( depth != tdepth && tdepth != std::max(CV_32F, depth) )
{
_templ.convertTo(templ, std::max(CV_32F, depth));
tdepth = templ.depth();
}
CV_Assert( depth == tdepth || tdepth == CV_32F);
CV_Assert( corrsize.height <= img.rows + templ.rows - 1 &&
corrsize.width <= img.cols + templ.cols - 1 );
CV_Assert( ccn == 1 || delta == 0 );
corr.create(corrsize, ctype);
int maxDepth = depth > CV_8S ? CV_64F : std::max(std::max(CV_32F, tdepth), cdepth);
Size blocksize, dftsize;
blocksize.width = cvRound(templ.cols*blockScale);
blocksize.width = std::max( blocksize.width, minBlockSize - templ.cols + 1 );
blocksize.width = std::min( blocksize.width, corr.cols );
blocksize.height = cvRound(templ.rows*blockScale);
blocksize.height = std::max( blocksize.height, minBlockSize - templ.rows + 1 );
blocksize.height = std::min( blocksize.height, corr.rows );
dftsize.width = std::max(getOptimalDFTSize(blocksize.width + templ.cols - 1), 2);
dftsize.height = getOptimalDFTSize(blocksize.height + templ.rows - 1);
if( dftsize.width <= 0 || dftsize.height <= 0 )
CV_Error( CV_StsOutOfRange, "the input arrays are too big" );
// recompute block size
blocksize.width = dftsize.width - templ.cols + 1;
blocksize.width = MIN( blocksize.width, corr.cols );
blocksize.height = dftsize.height - templ.rows + 1;
blocksize.height = MIN( blocksize.height, corr.rows );
Mat dftTempl( dftsize.height*tcn, dftsize.width, maxDepth );
Mat dftImg( dftsize, maxDepth );
int i, k, bufSize = 0;
if( tcn > 1 && tdepth != maxDepth )
bufSize = templ.cols*templ.rows*CV_ELEM_SIZE(tdepth);
if( cn > 1 && depth != maxDepth )
bufSize = std::max( bufSize, (blocksize.width + templ.cols - 1)*
(blocksize.height + templ.rows - 1)*CV_ELEM_SIZE(depth));
if( (ccn > 1 || cn > 1) && cdepth != maxDepth )
bufSize = std::max( bufSize, blocksize.width*blocksize.height*CV_ELEM_SIZE(cdepth));
buf.resize(bufSize);
// compute DFT of each template plane
for( k = 0; k < tcn; k++ )
{
int yofs = k*dftsize.height;
Mat src = templ;
Mat dst(dftTempl, Rect(0, yofs, dftsize.width, dftsize.height));
Mat dst1(dftTempl, Rect(0, yofs, templ.cols, templ.rows));
if( tcn > 1 )
{
src = tdepth == maxDepth ? dst1 : Mat(templ.size(), tdepth, &buf[0]);
int pairs[] = {k, 0};
mixChannels(&templ, 1, &src, 1, pairs, 1);
}
if( dst1.data != src.data )
src.convertTo(dst1, dst1.depth());
if( dst.cols > templ.cols )
{
Mat part(dst, Range(0, templ.rows), Range(templ.cols, dst.cols));
part = Scalar::all(0);
}
dft(dst, dst, 0, templ.rows);
}
int tileCountX = (corr.cols + blocksize.width - 1)/blocksize.width;
int tileCountY = (corr.rows + blocksize.height - 1)/blocksize.height;
int tileCount = tileCountX * tileCountY;
Size wholeSize = img.size();
Point roiofs(0,0);
Mat img0 = img;
if( !(borderType & BORDER_ISOLATED) )
{
img.locateROI(wholeSize, roiofs);
img0.adjustROI(roiofs.y, wholeSize.height-img.rows-roiofs.y,
roiofs.x, wholeSize.width-img.cols-roiofs.x);
}
borderType |= BORDER_ISOLATED;
// calculate correlation by blocks
for( i = 0; i < tileCount; i++ )
{
int x = (i%tileCountX)*blocksize.width;
int y = (i/tileCountX)*blocksize.height;
Size bsz(std::min(blocksize.width, corr.cols - x),
std::min(blocksize.height, corr.rows - y));
Size dsz(bsz.width + templ.cols - 1, bsz.height + templ.rows - 1);
int x0 = x - anchor.x + roiofs.x, y0 = y - anchor.y + roiofs.y;
int x1 = std::max(0, x0), y1 = std::max(0, y0);
int x2 = std::min(img0.cols, x0 + dsz.width);
int y2 = std::min(img0.rows, y0 + dsz.height);
Mat src0(img0, Range(y1, y2), Range(x1, x2));
Mat dst(dftImg, Rect(0, 0, dsz.width, dsz.height));
Mat dst1(dftImg, Rect(x1-x0, y1-y0, x2-x1, y2-y1));
Mat cdst(corr, Rect(x, y, bsz.width, bsz.height));
for( k = 0; k < cn; k++ )
{
Mat src = src0;
dftImg = Scalar::all(0);
if( cn > 1 )
{
src = depth == maxDepth ? dst1 : Mat(y2-y1, x2-x1, depth, &buf[0]);
int pairs[] = {k, 0};
mixChannels(&src0, 1, &src, 1, pairs, 1);
}
if( dst1.data != src.data )
src.convertTo(dst1, dst1.depth());
if( x2 - x1 < dsz.width || y2 - y1 < dsz.height )
copyMakeBorder(dst1, dst, y1-y0, dst.rows-dst1.rows-(y1-y0),
x1-x0, dst.cols-dst1.cols-(x1-x0), borderType);
dft( dftImg, dftImg, 0, dsz.height );
Mat dftTempl1(dftTempl, Rect(0, tcn > 1 ? k*dftsize.height : 0,
dftsize.width, dftsize.height));
mulSpectrums(dftImg, dftTempl1, dftImg, 0, true);
dft( dftImg, dftImg, DFT_INVERSE + DFT_SCALE, bsz.height );
src = dftImg(Rect(0, 0, bsz.width, bsz.height));
if( ccn > 1 )
{
if( cdepth != maxDepth )
{
Mat plane(bsz, cdepth, &buf[0]);
src.convertTo(plane, cdepth, 1, delta);
src = plane;
}
int pairs[] = {0, k};
mixChannels(&src, 1, &cdst, 1, pairs, 1);
}
else
{
if( k == 0 )
src.convertTo(cdst, cdepth, 1, delta);
else
{
if( maxDepth != cdepth )
{
Mat plane(bsz, cdepth, &buf[0]);
src.convertTo(plane, cdepth);
src = plane;
}
add(src, cdst, cdst);
}
}
}
}
}
int test_layer_channel_normalize()
{
// Blog: http://blog.csdn.net/fengbingchun/article/details/76976024
#ifdef __linux__
std::string image_name{ "test_data/images/lena.png" };
#else
std::string image_name{ "E:/GitCode/CUDA_Test/test_data/images/lena.png" };
#endif
cv::Mat matSrc = cv::imread(image_name);
if (!matSrc.data) {
fprintf(stderr, "read image fail: %s\n", image_name.c_str());
return -1;
}
const int width{ 511 }, height{ 473 }, channels{ 3 };
cv::resize(matSrc, matSrc, cv::Size(width, height));
matSrc.convertTo(matSrc, CV_32FC3);
std::vector<cv::Mat> matSplit;
cv::split(matSrc, matSplit);
CHECK(matSplit.size() == channels);
std::unique_ptr<float[]> data(new float[matSplit[0].cols * matSplit[0].rows * channels]);
size_t length{ matSplit[0].cols * matSplit[0].rows * sizeof(float) };
for (int i = 0; i < channels; ++i) {
memcpy(data.get() + matSplit[0].cols * matSplit[0].rows * i, matSplit[i].data, length);
}
float elapsed_time1{ 0.f }, elapsed_time2{ 0.f }; // milliseconds
std::unique_ptr<float[]> dst1(new float[matSplit[0].cols * matSplit[0].rows * channels]);
std::unique_ptr<float[]> dst2(new float[matSplit[0].cols * matSplit[0].rows * channels]);
int ret = layer_channel_normalize_cpu(data.get(), dst1.get(), width, height, channels, &elapsed_time1);
if (ret != 0) PRINT_ERROR_INFO(image_normalize_cpu);
ret = layer_channel_normalize_gpu(data.get(), dst2.get(), width, height, channels, &elapsed_time2);
if (ret != 0) PRINT_ERROR_INFO(image_normalize_gpu);
int count{ 0 }, num{ width * height * channels };
for (int i = 0; i < num; ++i) {
if (fabs(dst1[i] - dst2[i]) > 0.01/*EPS_*/) {
fprintf(stderr, "index: %d, val1: %f, val2: %f\n", i, dst1[i], dst2[i]);
++count;
}
if (count > 100) return -1;
}
std::vector<cv::Mat> merge(channels);
for (int i = 0; i < channels; ++i) {
merge[i] = cv::Mat(height, width, CV_32FC1, dst2.get() + i * width * height);
}
cv::Mat dst3;
cv::merge(merge, dst3);
dst3.convertTo(dst3, CV_8UC3, 255.f);
#ifdef __linux__
cv::imwrite("test_data/images/image_normalize.png", dst3);
#else
cv::imwrite("E:/GitCode/CUDA_Test/test_data/images/image_normalize.png", dst3);
#endif
//cv::resize(matSrc, matSrc, cv::Size(width, height));
#ifdef __linux__
//cv::imwrite("test_data/images/image_src.png", matSrc);
#else
//cv::imwrite("E:/GitCode/CUDA_Test/test_data/images/image_src.png", matSrc);
#endif
fprintf(stderr, "test layer channel normalize: cpu run time: %f ms, gpu run time: %f ms\n", elapsed_time1, elapsed_time2);
return 0;
}