Chiebot-Mirror
/
opencv
mirror of https://github.com/opencv/opencv.git
8
0
Fork 0
Code Issues Projects Releases Wiki Activity

Merge pull request #1878 from ilya-lavrenov:ocl_resize_area_master

Browse Source
pull/1895/merge
Andrey Pavlenko 11 years ago committed by OpenCV Buildbot
parent d83094b412 09795e3265
commit daefe6983f
8 changed files with 260 additions and 38 deletions
Whitespace
Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL
Split View
Diff Options
Show Stats Download Patch File Download Diff File
  1. 8
      modules/core/include/opencv2/core/ocl.hpp
  2. 16
      modules/core/src/ocl.cpp
  3. 8
      modules/core/test/test_umat.cpp
  4. 8
      modules/imgproc/src/color.cpp
  5. 142
      modules/imgproc/src/imgwarp.cpp
  6. 99
      modules/imgproc/src/opencl/resize.cl
  7. 15
      modules/imgproc/test/ocl/test_warp.cpp
  8. 2
      modules/ts/include/opencv2/ts/ocl_test.hpp

8
modules/core/include/opencv2/core/ocl.hpp
Unescape View File

@ -250,8 +250,12 @@ public:
KernelArg();
static KernelArg Local() { return KernelArg(LOCAL, 0); }
static KernelArg PtrOnly(const UMat & m)
{ return KernelArg(PTR_ONLY, (UMat*)&m); }
static KernelArg PtrWriteOnly(const UMat& m)
{ return KernelArg(PTR_ONLY+WRITE_ONLY, (UMat*)&m); }
static KernelArg PtrReadOnly(const UMat& m)
{ return KernelArg(PTR_ONLY+READ_ONLY, (UMat*)&m); }
static KernelArg PtrReadWrite(const UMat& m)
{ return KernelArg(PTR_ONLY+READ_WRITE, (UMat*)&m); }
static KernelArg ReadWrite(const UMat& m, int wscale=1)
{ return KernelArg(READ_WRITE, (UMat*)&m, wscale); }
static KernelArg ReadWriteNoSize(const UMat& m, int wscale=1)

16
modules/core/src/ocl.cpp
Unescape View File

@ -2197,10 +2197,10 @@ int Kernel::set(int i, const UMat& m)
int Kernel::set(int i, const KernelArg& arg)
{
CV_Assert( i >= 0 );
if( i == 0 )
p->cleanupUMats();
if( !p || !p->handle )
return -1;
if( i == 0 )
p->cleanupUMats();
if( arg.m )
{
int accessFlags = ((arg.flags & KernelArg::READ_ONLY) ? ACCESS_READ : 0) +
@ -2222,7 +2222,7 @@ int Kernel::set(int i, const KernelArg& arg)
{
int cols = u2d.cols*arg.wscale;
clSetKernelArg(p->handle, (cl_uint)i, sizeof(u2d.rows), &u2d.rows);
clSetKernelArg(p->handle, (cl_uint)(i+1), sizeof(u2d.cols), &cols);
clSetKernelArg(p->handle, (cl_uint)(i+1), sizeof(cols), &cols);
i += 2;
}
}
@ -2256,10 +2256,17 @@ bool Kernel::run(int dims, size_t globalsize[], size_t localsize[],
{
if(!p || !p->handle || p->e != 0)
return false;
AutoBuffer<size_t> _globalSize(dims);
size_t * globalSizePtr = (size_t *)_globalSize;
for (int i = 0; i < dims; ++i)
globalSizePtr[i] = localsize == NULL ? globalsize[i] :
((globalsize[i] + localsize[i] - 1) / localsize[i]) * localsize[i];
cl_command_queue qq = getQueue(q);
size_t offset[CV_MAX_DIM] = {0};
cl_int retval = clEnqueueNDRangeKernel(qq, p->handle, (cl_uint)dims,
offset, globalsize, localsize, 0, 0,
offset, globalSizePtr, localsize, 0, 0,
sync ? 0 : &p->e);
if( sync || retval < 0 )
{
@ -2350,6 +2357,7 @@ struct Program::Impl
void** deviceList = deviceListBuf;
for( i = 0; i < n; i++ )
deviceList[i] = ctx.device(i).ptr();
retval = clBuildProgram(handle, n,
(const cl_device_id*)deviceList,
buildflags.c_str(), 0, 0);

8
modules/core/test/test_umat.cpp
Unescape View File

@ -107,8 +107,8 @@ bool CV_UMatTest::TestUMat()
ra += Scalar::all(1.f);
{
Mat temp = ura.getMat(ACCESS_RW);
temp += Scalar::all(1.f);
Mat temp = ura.getMat(ACCESS_RW);
temp += Scalar::all(1.f);
}
ra.copyTo(rb);
CHECK_DIFF(ra, rb);
@ -146,8 +146,8 @@ bool CV_UMatTest::TestUMat()
CHECK_DIFF(rc0, rc);
{
UMat tmp = rc0.getUMat(ACCESS_WRITE);
cv::max(ura, urb, tmp);
UMat tmp = rc0.getUMat(ACCESS_WRITE);
cv::max(ura, urb, tmp);
}
CHECK_DIFF(rc0, rc);

8
modules/imgproc/src/color.cpp
Unescape View File

@ -2875,7 +2875,7 @@ static bool ocl_cvtColor( InputArray _src, OutputArray _dst, int code, int dcn )
k.create("RGB2XYZ", ocl::imgproc::cvtcolor_oclsrc,
format("-D depth=%d -D scn=%d -D dcn=3 -D bidx=%d", depth, scn, bidx));
k.args(ocl::KernelArg::ReadOnlyNoSize(src), ocl::KernelArg::WriteOnly(dst), ocl::KernelArg::PtrOnly(c));
k.args(ocl::KernelArg::ReadOnlyNoSize(src), ocl::KernelArg::WriteOnly(dst), ocl::KernelArg::PtrReadOnly(c));
return k.run(2, globalsize, 0, false);
}
case COLOR_XYZ2BGR: case COLOR_XYZ2RGB:
@ -2924,7 +2924,7 @@ static bool ocl_cvtColor( InputArray _src, OutputArray _dst, int code, int dcn )
k.create("XYZ2RGB", ocl::imgproc::cvtcolor_oclsrc,
format("-D depth=%d -D scn=3 -D dcn=%d -D bidx=%d", depth, dcn, bidx));
k.args(ocl::KernelArg::ReadOnlyNoSize(src), ocl::KernelArg::WriteOnly(dst), ocl::KernelArg::PtrOnly(c));
k.args(ocl::KernelArg::ReadOnlyNoSize(src), ocl::KernelArg::WriteOnly(dst), ocl::KernelArg::PtrReadOnly(c));
return k.run(2, globalsize, 0, false);
}
case COLOR_BGR2HSV: case COLOR_RGB2HSV: case COLOR_BGR2HSV_FULL: case COLOR_RGB2HSV_FULL:
@ -2980,8 +2980,8 @@ static bool ocl_cvtColor( InputArray _src, OutputArray _dst, int code, int dcn )
depth, hrange, bidx, scn));
k.args(ocl::KernelArg::ReadOnlyNoSize(src), ocl::KernelArg::WriteOnly(dst),
ocl::KernelArg::PtrOnly(sdiv_data), hrange == 256 ? ocl::KernelArg::PtrOnly(hdiv_data256) :
ocl::KernelArg::PtrOnly(hdiv_data180));
ocl::KernelArg::PtrReadOnly(sdiv_data), hrange == 256 ? ocl::KernelArg::PtrReadOnly(hdiv_data256) :
ocl::KernelArg::PtrReadOnly(hdiv_data180));
return k.run(2, globalsize, NULL, false);
}

142
modules/imgproc/src/imgwarp.cpp
Unescape View File

@ -48,8 +48,6 @@
#include "precomp.hpp"
#include "opencl_kernels.hpp"
#include <iostream>
#include <vector>
#if defined (HAVE_IPP) && (IPP_VERSION_MAJOR >= 7)
static IppStatus sts = ippInit();
@ -1902,18 +1900,72 @@ private:
};
#endif
static void ocl_computeResizeAreaTabs(int ssize, int dsize, double scale, int * const map_tab,
float * const alpha_tab, int * const ofs_tab)
{
int k = 0, dx = 0;
for ( ; dx < dsize; dx++)
{
ofs_tab[dx] = k;
double fsx1 = dx * scale;
double fsx2 = fsx1 + scale;
double cellWidth = std::min(scale, ssize - fsx1);
int sx1 = cvCeil(fsx1), sx2 = cvFloor(fsx2);
sx2 = std::min(sx2, ssize - 1);
sx1 = std::min(sx1, sx2);
if (sx1 - fsx1 > 1e-3)
{
map_tab[k] = sx1 - 1;
alpha_tab[k++] = (float)((sx1 - fsx1) / cellWidth);
}
for (int sx = sx1; sx < sx2; sx++)
{
map_tab[k] = sx;
alpha_tab[k++] = float(1.0 / cellWidth);
}
if (fsx2 - sx2 > 1e-3)
{
map_tab[k] = sx2;
alpha_tab[k++] = (float)(std::min(std::min(fsx2 - sx2, 1.), cellWidth) / cellWidth);
}
}
ofs_tab[dx] = k;
}
static void ocl_computeResizeAreaFastTabs(int * dmap_tab, int * smap_tab, int scale, int dcols, int scol)
{
for (int i = 0; i < dcols; ++i)
dmap_tab[i] = scale * i;
for (int i = 0, size = dcols * scale; i < size; ++i)
smap_tab[i] = std::min(scol - 1, i);
}
static bool ocl_resize( InputArray _src, OutputArray _dst, Size dsize,
double fx, double fy, int interpolation)
{
int type = _src.type(), depth = CV_MAT_DEPTH(type), cn = CV_MAT_CN(type);
if( !(cn <= 4 &&
(interpolation == INTER_NEAREST ||
(interpolation == INTER_LINEAR))) )
double inv_fx = 1. / fx, inv_fy = 1. / fy;
float inv_fxf = (float)inv_fx, inv_fyf = (float)inv_fy;
if( cn == 3 || !(cn <= 4 &&
(interpolation == INTER_NEAREST || interpolation == INTER_LINEAR ||
(interpolation == INTER_AREA && inv_fx >= 1 && inv_fy >= 1) )) )
return false;
UMat src = _src.getUMat();
_dst.create(dsize, type);
UMat dst = _dst.getUMat();
ocl::Kernel k;
size_t globalsize[] = { dst.cols, dst.rows };
if (interpolation == INTER_LINEAR)
{
@ -1929,14 +1981,86 @@ static bool ocl_resize( InputArray _src, OutputArray _dst, Size dsize,
else if (interpolation == INTER_NEAREST)
{
k.create("resizeNN", ocl::imgproc::resize_oclsrc,
format("-D INTER_NEAREST -D PIXTYPE=%s", ocl::memopTypeToStr(type) ));
format("-D INTER_NEAREST -D PIXTYPE=%s -D cn", ocl::memopTypeToStr(type), cn));
}
else if (interpolation == INTER_AREA)
{
int iscale_x = saturate_cast<int>(inv_fx);
int iscale_y = saturate_cast<int>(inv_fy);
bool is_area_fast = std::abs(inv_fx - iscale_x) < DBL_EPSILON &&
std::abs(inv_fy - iscale_y) < DBL_EPSILON;
int wdepth = std::max(depth, is_area_fast ? CV_32S : CV_32F);
int wtype = CV_MAKE_TYPE(wdepth, cn);
char cvt[2][40];
String buildOption = format("-D INTER_AREA -D T=%s -D WTV=%s -D convertToWTV=%s",
ocl::typeToStr(type), ocl::typeToStr(wtype),
ocl::convertTypeStr(depth, wdepth, cn, cvt[0]));
UMat alphaOcl, tabofsOcl, mapOcl;
UMat dmap, smap;
if (is_area_fast)
{
int wdepth2 = std::max(CV_32F, depth), wtype2 = CV_MAKE_TYPE(wdepth2, cn);
buildOption = buildOption + format(" -D convertToT=%s -D WT2V=%s -D convertToWT2V=%s -D INTER_AREA_FAST"
" -D XSCALE=%d -D YSCALE=%d -D SCALE=%f",
ocl::convertTypeStr(wdepth2, depth, cn, cvt[0]),
ocl::typeToStr(wtype2), ocl::convertTypeStr(wdepth, wdepth2, cn, cvt[1]),
iscale_x, iscale_y, 1.0f / (iscale_x * iscale_y));
k.create("resizeAREA_FAST", ocl::imgproc::resize_oclsrc, buildOption);
int smap_tab_size = dst.cols * iscale_x + dst.rows * iscale_y;
AutoBuffer<int> dmap_tab(dst.cols + dst.rows), smap_tab(smap_tab_size);
int * dxmap_tab = dmap_tab, * dymap_tab = dxmap_tab + dst.cols;
int * sxmap_tab = smap_tab, * symap_tab = smap_tab + dst.cols * iscale_y;
ocl_computeResizeAreaFastTabs(dxmap_tab, sxmap_tab, iscale_x, dst.cols, src.cols);
ocl_computeResizeAreaFastTabs(dymap_tab, symap_tab, iscale_y, dst.rows, src.rows);
Mat(1, dst.cols + dst.rows, CV_32SC1, (void *)dmap_tab).copyTo(dmap);
Mat(1, smap_tab_size, CV_32SC1, (void *)smap_tab).copyTo(smap);
}
else
{
buildOption = buildOption + format(" -D convertToT=%s", ocl::convertTypeStr(wdepth, depth, cn, cvt[0]));
k.create("resizeAREA", ocl::imgproc::resize_oclsrc, buildOption);
Size ssize = src.size();
int xytab_size = (ssize.width + ssize.height) << 1;
int tabofs_size = dsize.height + dsize.width + 2;
AutoBuffer<int> _xymap_tab(xytab_size), _xyofs_tab(tabofs_size);
AutoBuffer<float> _xyalpha_tab(xytab_size);
int * xmap_tab = _xymap_tab, * ymap_tab = _xymap_tab + (ssize.width << 1);
float * xalpha_tab = _xyalpha_tab, * yalpha_tab = _xyalpha_tab + (ssize.width << 1);
int * xofs_tab = _xyofs_tab, * yofs_tab = _xyofs_tab + dsize.width + 1;
ocl_computeResizeAreaTabs(ssize.width, dsize.width, inv_fx, xmap_tab, xalpha_tab, xofs_tab);
ocl_computeResizeAreaTabs(ssize.height, dsize.height, inv_fy, ymap_tab, yalpha_tab, yofs_tab);
// loading precomputed arrays to GPU
Mat(1, xytab_size, CV_32FC1, (void *)_xyalpha_tab).copyTo(alphaOcl);
Mat(1, xytab_size, CV_32SC1, (void *)_xymap_tab).copyTo(mapOcl);
Mat(1, tabofs_size, CV_32SC1, (void *)_xyofs_tab).copyTo(tabofsOcl);
}
ocl::KernelArg srcarg = ocl::KernelArg::ReadOnly(src), dstarg = ocl::KernelArg::WriteOnly(dst);
if (is_area_fast)
k.args(srcarg, dstarg, ocl::KernelArg::PtrReadOnly(dmap), ocl::KernelArg::PtrReadOnly(smap));
else
k.args(srcarg, dstarg, inv_fxf, inv_fyf, ocl::KernelArg::PtrReadOnly(tabofsOcl),
ocl::KernelArg::PtrReadOnly(mapOcl), ocl::KernelArg::PtrReadOnly(alphaOcl));
return k.run(2, globalsize, NULL, false);
}
if( k.empty() )
return false;
k.args(ocl::KernelArg::ReadOnly(src), ocl::KernelArg::WriteOnly(dst),
(float)(1./fx), (float)(1./fy));
size_t globalsize[] = { dst.cols, dst.rows };
(float)inv_fx, (float)inv_fy);
return k.run(2, globalsize, 0, false);
}
@ -2069,7 +2193,7 @@ void cv::resize( InputArray _src, OutputArray _dst, Size dsize,
}
if( ocl::useOpenCL() && _dst.kind() == _InputArray::UMAT &&
ocl_resize(_src, _dst, dsize, inv_scale_x, inv_scale_y, interpolation) )
ocl_resize(_src, _dst, dsize, inv_scale_x, inv_scale_y, interpolation))
return;
Mat src = _src.getMat();

99
modules/imgproc/src/opencl/resize.cl
Unescape View File

@ -43,16 +43,8 @@
//
//M*/
// resize kernel
// Currently, CV_8UC1 CV_8UC4 CV_32FC1 and CV_32FC4are supported.
// We shall support other types later if necessary.
#if defined DOUBLE_SUPPORT
#pragma OPENCL EXTENSION cl_khr_fp64:enable
#define F double
#else
#define F float
#endif
#define INTER_RESIZE_COEF_BITS 11
@ -141,8 +133,8 @@ __kernel void resizeNN(__global const uchar* srcptr, int srcstep, int srcoffset,
if( dx < dstcols && dy < dstrows )
{
F s1 = dx*ifx;
F s2 = dy*ify;
float s1 = dx*ifx;
float s2 = dy*ify;
int sx = min(convert_int_rtz(s1), srccols-1);
int sy = min(convert_int_rtz(s2), srcrows-1);
@ -153,4 +145,91 @@ __kernel void resizeNN(__global const uchar* srcptr, int srcstep, int srcoffset,
}
}
#elif defined INTER_AREA
#define TSIZE ((int)(sizeof(T)))
#ifdef INTER_AREA_FAST
__kernel void resizeAREA_FAST(__global const uchar * src, int src_step, int src_offset, int src_rows, int src_cols,
__global uchar * dst, int dst_step, int dst_offset, int dst_rows, int dst_cols,
__global const int * dmap_tab, __global const int * smap_tab)
{
int dx = get_global_id(0);
int dy = get_global_id(1);
if (dx < dst_cols && dy < dst_rows)
{
int dst_index = mad24(dy, dst_step, dst_offset);
__global const int * xmap_tab = dmap_tab;
__global const int * ymap_tab = dmap_tab + dst_cols;
__global const int * sxmap_tab = smap_tab;
__global const int * symap_tab = smap_tab + XSCALE * dst_cols;
int sx = xmap_tab[dx], sy = ymap_tab[dy];
WTV sum = (WTV)(0);
#pragma unroll
for (int y = 0; y < YSCALE; ++y)
{
int src_index = mad24(symap_tab[y + sy], src_step, src_offset);
#pragma unroll
for (int x = 0; x < XSCALE; ++x)
sum += convertToWTV(((__global const T*)(src + src_index))[sxmap_tab[sx + x]]);
}
((__global T*)(dst + dst_index))[dx] = convertToT(convertToWT2V(sum) * (WT2V)(SCALE));
}
}
#else
__kernel void resizeAREA(__global const uchar * src, int src_step, int src_offset, int src_rows, int src_cols,
__global uchar * dst, int dst_step, int dst_offset, int dst_rows, int dst_cols,
float ifx, float ify, __global const int * ofs_tab,
__global const int * map_tab, __global const float * alpha_tab)
{
int dx = get_global_id(0);
int dy = get_global_id(1);
if (dx < dst_cols && dy < dst_rows)
{
int dst_index = mad24(dy, dst_step, dst_offset);
__global const int * xmap_tab = map_tab;
__global const int * ymap_tab = (__global const int *)(map_tab + (src_cols << 1));
__global const float * xalpha_tab = alpha_tab;
__global const float * yalpha_tab = (__global const float *)(alpha_tab + (src_cols << 1));
__global const int * xofs_tab = ofs_tab;
__global const int * yofs_tab = (__global const int *)(ofs_tab + dst_cols + 1);
int xk0 = xofs_tab[dx], xk1 = xofs_tab[dx + 1];
int yk0 = yofs_tab[dy], yk1 = yofs_tab[dy + 1];
int sy0 = ymap_tab[yk0], sy1 = ymap_tab[yk1 - 1];
int sx0 = xmap_tab[xk0], sx1 = xmap_tab[xk1 - 1];
WTV sum = (WTV)(0), buf;
int src_index = mad24(sy0, src_step, src_offset);
for (int sy = sy0, yk = yk0; sy <= sy1; ++sy, src_index += src_step, ++yk)
{
WTV beta = (WTV)(yalpha_tab[yk]);
buf = (WTV)(0);
for (int sx = sx0, xk = xk0; sx <= sx1; ++sx, ++xk)
{
WTV alpha = (WTV)(xalpha_tab[xk]);
buf += convertToWTV(((__global const T*)(src + src_index))[sx]) * alpha;
}
sum += buf * beta;
}
((__global T*)(dst + dst_index))[dx] = convertToT(sum);
}
}
#endif
#endif

15
modules/imgproc/test/ocl/test_warp.cpp
Unescape View File

@ -127,13 +127,20 @@ OCL_TEST_P(Resize, Mat)
/////////////////////////////////////////////////////////////////////////////////////
OCL_INSTANTIATE_TEST_CASE_P(ImgprocWarp, Resize, Combine(
Values(CV_8UC1, CV_8UC4, CV_16UC2, CV_32FC1, CV_32FC4),
Values(0.5, 1.5, 2.0),
Values(0.5, 1.5, 2.0),
OCL_INSTANTIATE_TEST_CASE_P(ImgprocWarpResize, Resize, Combine(
Values((MatType)CV_8UC1, CV_8UC4, CV_32FC1, CV_32FC4),
Values(0.7, 0.4, 2.0),
Values(0.3, 0.6, 2.0),
Values((Interpolation)INTER_NEAREST, (Interpolation)INTER_LINEAR),
Bool()));
OCL_INSTANTIATE_TEST_CASE_P(ImgprocWarpResizeArea, Resize, Combine(
Values((MatType)CV_8UC1, CV_8UC4, CV_32FC1, CV_32FC4),
Values(0.7, 0.4, 0.5),
Values(0.3, 0.6, 0.5),
Values((Interpolation)INTER_AREA),
Bool()));
} } // namespace cvtest::ocl
#endif // HAVE_OPENCL

2
modules/ts/include/opencv2/ts/ocl_test.hpp
Unescape View File

@ -305,7 +305,7 @@ IMPLEMENT_PARAM_CLASS(Channels, int)
#define OCL_ALL_DEPTHS Values(CV_8U, CV_8S, CV_16U, CV_16S, CV_32S, CV_32F)
#define OCL_ALL_CHANNELS Values(1, 2, 3, 4)
CV_ENUM(Interpolation, INTER_NEAREST, INTER_LINEAR, INTER_CUBIC)
CV_ENUM(Interpolation, INTER_NEAREST, INTER_LINEAR, INTER_CUBIC, INTER_AREA)
#define OCL_INSTANTIATE_TEST_CASE_P(prefix, test_case_name, generator) \
INSTANTIATE_TEST_CASE_P(OCL_ ## prefix, test_case_name, generator)

Write Preview
Loading…
Cancel
Save