mirror/opencv
1
0
Fork 0
mirror of https://github.com/opencv/opencv.git synced 2024-12-14 08:59:11 +08:00
Code Issues Actions Packages Projects Releases Wiki Activity
16,323 Commits 6 Branches 127 Tags 3 GiB
3c85200989
Commit Graph

253 Commits

Author SHA1 Message Date
Zhigang Gong
3c85200989 Avoid negative index for a local buffer in Canny.cl. 
int pix_per_thr = l_counter / LOCAL_TOTAL + ((lid < mod) ? 1 : 0);
The pix_per_thr * LOCAL_TOTAL may be larger than l_counter.
Thus the index of l_stack may be negative which may cause serious
problems. Let's skip the loop when we get negative index and we need
to add back the lcounter to keep its balance and avoid potential
negative counter.

Signed-off-by: Zhigang Gong <zhigang.gong@intel.com>
 2015-05-26 08:48:05 +08:00
Pavel Rojtberg
1ea41e7246 fix gftt opencv kernel when using mask 2015-04-22 16:13:50 +02:00
Yan Wang
6e7050555e Optimize pyrUp_unrolled() by mad function. 
It could improve performance when image size is large.
E.g. OCL_PyrUpFixture_PyrUp.PyrUp/18
 2014-11-26 16:55:08 +08:00
Alexander Alekhin
569a95e9e1 Merge pull request #3394 from akarsakov:ocl_canny 2014-11-07 12:42:24 +00:00
Alexander Karsakov
7c870014fb Correctly unrolled some cycles 2014-11-07 12:13:00 +03:00
Alexander Karsakov
0ec0aeb7d0 Minor optimization for ocl_canny 2014-11-06 13:07:33 +03:00
vbystricky
957e5ef8eb Fix OpenCL version of HoughLinesP function 2014-11-05 14:31:06 +04:00
Alexander Karsakov
643c906f3d Added optimized loading to YUV2RGB_422 kernel 2014-10-28 15:07:51 +03:00
Alexander Karsakov
1466621f99 Added loading 4 pixels in line instead of 2 to RGB[A] -> YUV(420) kernel 2014-10-27 16:00:34 +03:00
Alexander Karsakov
60367907fe Used direct float calculations 2014-10-21 17:18:03 +03:00
Alexander Karsakov
5aa9ac9a77 Added OCL code for YUV422 -> RGB[A]|BGR[A] color conversion 2014-10-21 17:18:03 +03:00
Alexander Karsakov
c8707b891b Added OCL code for RGB[A]|BGR[A] -> YUV_[YV12|IYUV] color conversion 2014-10-21 17:18:03 +03:00
Alexander Karsakov
1cc17a7186 Added OCL code for YUV2BGR_YV12 and YUV2BGR_IYUV color conversions 2014-10-21 17:18:02 +03:00
Alexander Karsakov
85b60ee3cb Added support for YUV2RGB[A]_NV21 and YUV2BGR[A]_NV21 conversion 2014-10-21 17:18:02 +03:00
Vadim Pisarevsky
397870d7a5 Merge pull request #3279 from akarsakov:ocl_houghlines 2014-10-09 14:56:45 +00:00
Alexander Karsakov
66a8acfd3d Optimization for HoughLinesP 2014-10-07 17:53:33 +04:00
Alexander Alekhin
14d5358982 Merge pull request #3210 from akarsakov:ocl_gftt_opt 2014-10-07 09:06:54 +00:00
Alexander Karsakov
eaf5a163b1 Added HoughLinesP OCL implementation 2014-09-29 16:48:16 +04:00
Alexander Karsakov
3695a31606 Combined counter and corner buffers into one 2014-09-29 11:10:57 +04:00
Vadim Pisarevsky
470f427a95 Merge pull request #3232 from Chuanbo-Weng:master 2014-09-18 11:48:29 +00:00
Chuanbo Weng
c5552788c5 Use vload to read unaligned data instead of dereference operator. 
According to opencl 1.2 spec 6.1.5:
    For arguments to a __kernel function declared to be a pointer to a
    data type, the OpenCL compiler can assume that the pointee is always
    appropriately aligned as required by the data type. The behavior of
    an unaligned load or store is undefined, except for the
    vloadn, vload_halfn, vstoren, and vstore_halfn functions defined in
    section 6.12.7.

Original code read data of type T from address not aligned by multiple
of sizeof(T), so the result is incorrect. With this patch, the cases
./opencv_perf_imgproc
--gtest_filter=OCL_ImgSize_TmplSize_Method_MatType_MatchTemplate.MatchTemplate/*
could work well with beignet 0.9.3.

Signed-off-by: Chuanbo Weng <chuanbo.weng@intel.com>
 2014-09-17 19:28:07 +08:00
Alexander Karsakov
8c08714b8c Remove two "set" kernel call 2014-09-11 18:11:23 +04:00
vbystricky
b0bf8478e5 Optimization OpenCL version of Filter2D 2014-09-11 12:59:51 +04:00
Alexander Karsakov
39b27a19be Refactoring and optimization 2014-09-05 12:20:29 +04:00
Alexander Karsakov
d59a6fa518 Optimization for getLines 2014-09-05 11:37:16 +04:00
Alexander Karsakov
fee8f29f48 Refactoring, minor optimization 2014-09-04 16:31:30 +04:00
Alexander Karsakov
07d57db91c Fixed calculation of l_stack_size 2014-09-03 17:40:17 +04:00
Alexander Karsakov
214dab39f6 Fixed BORDER_REFLECT and BORDER_REFLECT_101 extrapolation for case x > 2*maxV 2014-09-02 11:53:31 +04:00
Alexander Alekhin
b332152bef Merge pull request #2956 from ilya-lavrenov:tapi_accumulate 2014-08-28 09:08:51 +00:00
Alexander Karsakov
f7aadd07f6 Added getLines, fill_accum_local kernels 2014-08-27 17:57:22 +04:00
Vadim Pisarevsky
d66815978a Merge pull request #3117 from KruchDmitriy:canny_opt 2014-08-27 10:07:55 +00:00
VBystricky
9ee0789174 Fix issues 2014-08-26 14:39:11 +04:00
vbystricky
e75cd74f5a Optimize OpenCL version of Laplacian filter for kernel size great than 3 2014-08-25 17:56:09 +04:00
Alexander Karsakov
038bfb98ec Added fill_accum kernel 2014-08-25 13:55:09 +04:00
Ilya Lavrenov
a350b76738 optimization of cv::accumulate** 2014-08-25 11:25:01 +04:00
Alexander Karsakov
5c1f71de51 Added make_point_list kernel 2014-08-22 16:50:01 +04:00
Vadim Pisarevsky
e7539bd2c8 Merge pull request #3144 from ElenaGvozdeva:ocl_morphSmall 2014-08-22 12:14:06 +00:00
U-KruchininD-ПК\KruchininD
6ed168d3af New optimization for canny 
new hysteresis

delete whitespaces

fix problem with mad24

Dynamic work group size

dynamic work group size

Fix problem with warnings

Fix some problems with border

Another one fix

Delete trailing whitespaces

some changes

fix problem with warning
 2014-08-22 11:22:15 +04:00
Alexander Karsakov
3d222d313b Fixed range for 'v' channel for 8U images 2014-08-21 17:22:06 +04:00
Elena Gvozdeva
5302e56071 fix for ocl_morphSmall 2014-08-21 16:31:24 +04:00
Vadim Pisarevsky
e9729a9601 multiple yet minor fixes to make most of the tests pass on Mac with Iris graphics 2014-08-16 00:29:10 +04:00
Alexander Karsakov
5898dcae4a Added ROUNDING_EPS for identical rounding after dividing on different platforms 2014-08-12 14:28:48 +04:00
Alexander Alekhin
d0f789dc90 Merge pull request #3044 from akarsakov:fix_ocl_tests 2014-08-07 20:14:17 +00:00
Alexander Karsakov
44fbfb2cf6 Fixed extrapolation in pyrDown 2014-08-07 10:39:25 +04:00
Alexander Karsakov
2a0b39d30a Fixed calculate_histogram kernel 2014-08-07 10:39:24 +04:00
Alexander Karsakov
eb9fdb0164 Fixed rounding in remap INTER_LINEAR mode 2014-08-07 10:39:24 +04:00
Alexander Karsakov
fec21239c8 Revert optimization for warpAffine INTER_NEAREST mode 2014-08-07 10:39:18 +04:00
Vadim Pisarevsky
bab5700a09 Merge pull request #3013 from ElenaGvozdeva:ocl_matchTemplate 2014-08-06 11:26:07 +00:00
Elena Gvozdeva
7999fbf765 fixed ocl_integral 2014-08-05 12:01:28 +04:00
Vadim Pisarevsky
d3cdfefa03 Merge pull request #3027 from vbystricky:cvtColorLuv2RGB 2014-08-02 21:52:24 +00:00