Commit Graph

62 Commits

Author SHA1 Message Date
Vladislav Vinogradov
bfc27271e2 Revert "disable gpu Canny and HoughCircles perf tests:"
This reverts commit dbce90692a.
2014-02-07 14:55:06 +04:00
Roman Donchenko
2bef73c968 Merge pull request #2249 from jet47:gpu-tests-disable 2014-01-31 20:49:05 +04:00
Vladislav Vinogradov
dbce90692a disable gpu Canny and HoughCircles perf tests:
it fails because driver terminates CUDA kernels after time out
2014-01-31 16:14:07 +04:00
Vladislav Vinogradov
8401ed9395 disable some gpu tests if library was built without CUFFT 2014-01-31 15:40:59 +04:00
Vladislav Vinogradov
063d8b4211 disable performance test for gpu generalized hough 2014-01-31 12:40:40 +04:00
Ilya Lavrenov
6b9ebcbf3d deleted extra semicolons 2014-01-07 02:52:30 +04:00
Vladislav Vinogradov
d13ecd0d76 increase epsilon for GEMM & Convolve sanity tests for CUDA 5.5 2013-08-02 17:31:34 +04:00
Vladislav Vinogradov
c8d1fccdac compare cv::gpu::calcHist with cv::calcHist 2013-08-01 18:00:05 +04:00
Andrey Kamaev
96b008cd29 Eliminate the need of ::testsing::ValuesIn() for CV_ENUM
Also cv::, cv::gpu:: and cv::ocl:: namespace prefixes can be safely omitted
inside CV_ENUM and CV_FLAGS
2013-04-15 19:39:49 +04:00
Vladislav Vinogradov
4d23e2c8c9 GPU implementation of CLAHE 2013-03-25 17:45:13 +04:00
Vladislav Vinogradov
f0b19d4659 updated license header in whole gpu module 2013-03-22 14:03:15 +04:00
Vladislav Vinogradov
b07bce11ef increased epsilon in sanity test 2013-03-21 13:40:07 +04:00
Vladislav Vinogradov
83def21ba9 increased time limits in perf tests 2013-03-21 13:40:07 +04:00
Vladislav Vinogradov
4a237af814 implemented Malvar, He, and Cutler Bayer Demosaicing on gpu 2013-03-14 18:11:12 +04:00
Vladislav Vinogradov
4ddf634c30 gpu : implement Bayer* -> Gray color conversion 2013-03-11 15:41:50 +04:00
Vladislav Vinogradov
f533f0dca9 fixed gpu sanity tests 2013-02-28 16:54:37 +04:00
Vladislav Vinogradov
a138e12a26 refactored gpu perf tests and fixed sanity tests 2013-02-26 13:53:38 +04:00
Vladislav Vinogradov
fe2e89df1b added gpu::HoughLinesP function 2013-02-13 15:56:58 +04:00
Vladislav Vinogradov
9cb4292d5c implemented Luv/Lab <-> RGB conversion 2013-01-23 21:05:08 +04:00
Vladislav Vinogradov
b7e6b5af1b fixed tests (call resetDevice, if there was a gpu failure) 2013-01-23 21:05:04 +04:00
OpenCV Buildbot
04384a71e4 Normalize line endings and whitespace 2012-10-17 15:32:23 +04:00
marina.kolpakova
d7ef6f078f temporarily disable HoughLines test 2012-10-10 16:13:25 +04:00
marina.kolpakova
1e344d7eba add sanity checks into the imgproc gpu performance tests 2012-10-10 16:13:24 +04:00
Anatoly Baksheev
5b31c46c4f added MatCn enum to GPU perf tests 2012-10-08 19:57:20 +04:00
Vadim Pisarevsky
06a13d9ba0 replaced jpegs with png's. Accuracy tests pass; performance tests were not tried yet 2012-10-02 23:07:46 +04:00
Andrey Kamaev
e73089bdb5 Fix windows build warnings 2012-09-17 13:30:28 +04:00
Vladislav Vinogradov
fd77ef7ece optimized gpu pyrDown and sepFilter2D 2012-09-13 10:21:43 +04:00
Vladislav Vinogradov
98c92f196e added Generalized Hough implementation 2012-09-10 16:49:40 +04:00
Anatoly Baksheev
0ba01afd83 added GPU bilateral filter + tests
added GPU non local means brute force filter + tests
2012-08-29 19:20:32 +04:00
Vadim Pisarevsky
633a8bfacc fixed many warnings (modified pull request 13) 2012-08-28 18:15:14 +04:00
Vladislav Vinogradov
c3f277b7bc gpu version of HoughCircles 2012-08-23 17:12:43 +04:00
Andrey Kamaev
f2a02feffa Fix Windows build issues
* No /arch:SSE3 in MSVC
* Warnings fixed or suppressed
2012-08-22 11:21:09 +04:00
Vladislav Vinogradov
c26d543e1e gpu::HoughLines : minor code improvements 2012-08-20 16:03:01 +04:00
Vladislav Vinogradov
a507d564a4 merged gpu_perf_cpu into gpu_perf 2012-08-17 17:34:58 +04:00
Vladislav Vinogradov
6e4eb722a6 updated gpu performance tests
now it executes only on one device
added posibility to specify device on which tests will be executed
2012-08-17 15:14:14 +04:00
Vladislav Vinogradov
c3fa7974e6 new optimized version 2012-08-14 17:00:57 +04:00
Vladislav Vinogradov
9201db32eb fixed overflow bugs, updated perf tests 2012-08-14 12:33:47 +04:00
Vladislav Vinogradov
1eefc69927 added accuracy and performance tests 2012-08-14 11:36:32 +04:00
Vladislav Vinogradov
8624d18ca5 new implementation of gpu debayer
* without border extrapolation
* with aligned write
* process 4 pixels per thread in 8u case
2012-08-07 13:34:11 +04:00
Vladislav Vinogradov
5ce896d9ee added perf tests for new gpu::cvtColor functions 2012-08-07 13:34:11 +04:00
Marina Kolpakova
e7f6c4b7ef scan operations are moved in separate header 2012-06-20 05:41:16 +00:00
Marina Kolpakova
f2c30cd90d resize area for big downscaling integration 2012-06-09 15:24:01 +00:00
Marina Kolpakova
16ef01efda set to linear interpolation type for upscaling with area interplation flag 2012-06-06 16:05:31 +00:00
Marina Kolpakova
83390ea4e1 performance test for INTER_AREA resize has been added. 2012-06-06 14:45:33 +00:00
Vladislav Vinogradov
f07250acf3 updated gpu performance tests 2012-05-23 12:58:01 +00:00
Anatoly Baksheev
671233cd46 gpu: added Cascade and mulAndScaleSpectrums perf tests 2012-05-22 18:58:01 +00:00
Vladislav Vinogradov
ade7394e77 refactored and fixed bugs in gpu warp functions (remap, resize, warpAffine, warpPerspective)
wrote more complicated tests for them
implemented own version of warpAffine and warpPerspective for different border interpolation types
refactored some gpu tests
2012-03-14 15:54:17 +00:00
Vladislav Vinogradov
e7dda44a07 updated image for StereoConstantSpaceBP regression test
updated gpu tests for CornerHarris and CornerMinEigen
moved direct convolution implementation to gpu::filter2D, gpu::convolve now use only DFT-based algorithm (Bug #1639)
2012-03-07 09:49:24 +00:00
Vladislav Vinogradov
2d30480982 created wrappers for new NPP functions
removed void integral(const GpuMat& src, GpuMat& sum, GpuMat& sqsum, Stream& stream) - it fails with NPP_NOT_IMPLEMENTED error
updated docs, accuracy and performance tests
2012-02-22 10:00:53 +00:00
Andrey Kamaev
65f5343ed5 Refactored performance tests. TEST_CYCLE macro is renamed to TEST_CYCLE_N; SIMPLE_TEST_CYCLE is renamed to TEST_CYCLE; from now 100 iterations are default for performance tests 2011-12-29 16:46:16 +00:00