potok/opencv
1
0
Fork 0
mirror of https://github.com/opencv/opencv.git synced 2026-07-30 15:53:03 +04:00
Code Packages Releases Wiki Activity

imgproc: align GaussianBlur/sepFilter2D OpenCL with CPU version

Browse Source
This commit is contained in:
Alexander Alekhin
2020-06-26 22:31:02 +00:00
parent 2fed41dfa5
commit 762a5c8334
7 changed files with 273 additions and 108 deletions
Download Patch File Download Diff File Expand all files Collapse all files
modules/imgproc/src/smooth.dispatch.cpp
+44 -24
View File
@@ -48,6 +48,7 @@
#include <opencv2/core/utils/configuration.private.hpp>
#include <vector>
#include <iostream>
#include "opencv2/core/hal/intrin.hpp"
#include "opencl_kernels_imgproc.hpp"
@@ -637,10 +638,9 @@ void GaussianBlur(InputArray _src, OutputArray _dst, Size ksize,
return;
}
bool useOpenCL = (ocl::isOpenCLActivated() && _dst.isUMat() && _src.dims() <= 2 &&
((ksize.width == 3 && ksize.height == 3) ||
(ksize.width == 5 && ksize.height == 5)) &&
_src.rows() > ksize.height && _src.cols() > ksize.width);
bool useOpenCL = ocl::isOpenCLActivated() && _dst.isUMat() && _src.dims() <= 2 &&
_src.rows() >= ksize.height && _src.cols() >= ksize.width &&
ksize.width > 1 && ksize.height > 1;
CV_UNUSED(useOpenCL);
int sdepth = CV_MAT_DEPTH(type), cn = CV_MAT_CN(type);
@@ -648,27 +648,13 @@ void GaussianBlur(InputArray _src, OutputArray _dst, Size ksize,
Mat kx, ky;
createGaussianKernels(kx, ky, type, ksize, sigma1, sigma2);
CV_OCL_RUN(useOpenCL, ocl_GaussianBlur_8UC1(_src, _dst, ksize, CV_MAT_DEPTH(type), kx, ky, borderType));
CV_OCL_RUN(useOpenCL && sdepth == CV_8U &&
((ksize.width == 3 && ksize.height == 3) ||
(ksize.width == 5 && ksize.height == 5)),
ocl_GaussianBlur_8UC1(_src, _dst, ksize, CV_MAT_DEPTH(type), kx, ky, borderType)
);
CV_OCL_RUN(_dst.isUMat() && _src.dims() <= 2 && (size_t)_src.rows() > kx.total() && (size_t)_src.cols() > kx.total(),
ocl_sepFilter2D(_src, _dst, sdepth, kx, ky, Point(-1, -1), 0, borderType))
Mat src = _src.getMat();
Mat dst = _dst.getMat();
Point ofs;
Size wsz(src.cols, src.rows);
if(!(borderType & BORDER_ISOLATED))
src.locateROI( wsz, ofs );
CALL_HAL(gaussianBlur, cv_hal_gaussianBlur, src.ptr(), src.step, dst.ptr(), dst.step, src.cols, src.rows, sdepth, cn,
ofs.x, ofs.y, wsz.width - src.cols - ofs.x, wsz.height - src.rows - ofs.y, ksize.width, ksize.height,
sigma1, sigma2, borderType&~BORDER_ISOLATED);
CV_OVX_RUN(true,
openvx_gaussianBlur(src, dst, ksize, sigma1, sigma2, borderType))
if(sdepth == CV_8U && ((borderType & BORDER_ISOLATED) || !_src.getMat().isSubmatrix()))
if(sdepth == CV_8U && ((borderType & BORDER_ISOLATED) || !_src.isSubmatrix()))
{
std::vector<ufixedpoint16> fkx, fky;
createGaussianKernels(fkx, fky, type, ksize, sigma1, sigma2);
@@ -684,6 +670,17 @@ void GaussianBlur(InputArray _src, OutputArray _dst, Size ksize,
}
else
{
CV_OCL_RUN(useOpenCL,
ocl_sepFilter2D_BitExact(_src, _dst, sdepth,
ksize,
(const uint16_t*)&fkx[0], (const uint16_t*)&fky[0],
Point(-1, -1), 0, borderType,
8/*shift_bits*/)
);
Mat src = _src.getMat();
Mat dst = _dst.getMat();
if (src.data == dst.data)
src = src.clone();
CV_CPU_DISPATCH(GaussianBlurFixedPoint, (src, dst, (const uint16_t*)&fkx[0], (int)fkx.size(), (const uint16_t*)&fky[0], (int)fky.size(), borderType),
@@ -692,6 +689,29 @@ void GaussianBlur(InputArray _src, OutputArray _dst, Size ksize,
}
}
#ifdef HAVE_OPENCL
if (useOpenCL)
{
sepFilter2D(_src, _dst, sdepth, kx, ky, Point(-1, -1), 0, borderType);
return;
}
#endif
Mat src = _src.getMat();
Mat dst = _dst.getMat();
Point ofs;
Size wsz(src.cols, src.rows);
if(!(borderType & BORDER_ISOLATED))
src.locateROI( wsz, ofs );
CALL_HAL(gaussianBlur, cv_hal_gaussianBlur, src.ptr(), src.step, dst.ptr(), dst.step, src.cols, src.rows, sdepth, cn,
ofs.x, ofs.y, wsz.width - src.cols - ofs.x, wsz.height - src.rows - ofs.y, ksize.width, ksize.height,
sigma1, sigma2, borderType&~BORDER_ISOLATED);
CV_OVX_RUN(true,
openvx_gaussianBlur(src, dst, ksize, sigma1, sigma2, borderType))
#if defined ENABLE_IPP_GAUSSIAN_BLUR
// IPP is not bit-exact to OpenCV implementation
CV_IPP_RUN_FAST(ipp_GaussianBlur(src, dst, ksize, sigma1, sigma2, borderType));