Alexander Karsakov [Thu, 6 Nov 2014 13:23:02 +0000 (16:23 +0300)]
Removed incorrect using of rootn() and powr() in ocl_pow
Vadim Pisarevsky [Thu, 6 Nov 2014 08:40:35 +0000 (08:40 +0000)]
Merge pull request #3372 from pokeefe:add-test-for-projection-decomposition
Vadim Pisarevsky [Thu, 6 Nov 2014 08:37:34 +0000 (08:37 +0000)]
Merge pull request #3391 from alalek:ocl_failure_messages
Vadim Pisarevsky [Thu, 6 Nov 2014 08:36:23 +0000 (08:36 +0000)]
Merge pull request #3360 from mirab:threshold_triangle
Alexander Alekhin [Wed, 5 Nov 2014 15:41:39 +0000 (18:41 +0300)]
ocl: change processing of OpenCL failures
disable "unwanted" messages
Miroslav Benes [Wed, 5 Nov 2014 15:04:26 +0000 (16:04 +0100)]
Make Otsu and Triangle algorithms mutually exclusive
Make Otsu and Triangle algorithms mutually exclusive in cv::threshold code.
Signed-off-by: Miroslav Benes <mbenes@suse.cz>
Vadim Pisarevsky [Wed, 5 Nov 2014 14:58:27 +0000 (14:58 +0000)]
Merge pull request #3389 from blackibiza84:master
Vadim Pisarevsky [Wed, 5 Nov 2014 14:57:45 +0000 (14:57 +0000)]
Merge pull request #3217 from avdmitry:samples_cpp_data
Vadim Pisarevsky [Wed, 5 Nov 2014 14:56:22 +0000 (14:56 +0000)]
Merge pull request #3377 from jlblancoc:master
Vadim Pisarevsky [Wed, 5 Nov 2014 11:42:51 +0000 (11:42 +0000)]
Merge pull request #3387 from YusukeSuzuki:cap_v4l_rgb24fix
Vadim Pisarevsky [Wed, 5 Nov 2014 11:41:12 +0000 (11:41 +0000)]
Merge pull request #3390 from vbystricky:fix_houghlines
Vadim Pisarevsky [Wed, 5 Nov 2014 11:26:52 +0000 (11:26 +0000)]
Merge pull request #3384 from ilya-lavrenov:neon_new
Vadim Pisarevsky [Wed, 5 Nov 2014 11:14:59 +0000 (11:14 +0000)]
Merge pull request #3373 from berak:patch-2
Vadim Pisarevsky [Wed, 5 Nov 2014 11:04:06 +0000 (11:04 +0000)]
Merge pull request #3388 from bartaandras:master
Michele Adduci [Wed, 5 Nov 2014 11:03:28 +0000 (12:03 +0100)]
Updated test_features2d.cpp with latest API change
Michele Adduci [Wed, 5 Nov 2014 11:00:31 +0000 (12:00 +0100)]
Updated ORB test with parameters
Vadim Pisarevsky [Wed, 5 Nov 2014 10:41:32 +0000 (10:41 +0000)]
Merge pull request #3315 from abak:seamless-refactor
vbystricky [Wed, 5 Nov 2014 10:31:06 +0000 (14:31 +0400)]
Fix OpenCL version of HoughLinesP function
Michele Adduci [Wed, 5 Nov 2014 09:42:02 +0000 (10:42 +0100)]
Fix cudafeatures2d and cudastereo with new 3.0 APIs for ORB and StereoBM
bartaandras [Tue, 4 Nov 2014 10:35:45 +0000 (11:35 +0100)]
Fixed calc_output_scale with NO_OUTPUT_SCALE flag set.
Yusuke Suzuki [Mon, 3 Nov 2014 08:48:10 +0000 (17:48 +0900)]
add V4L2_PIX_FMT_RGB24 convert code
berak [Sun, 2 Nov 2014 07:16:29 +0000 (08:16 +0100)]
support older gcc versions
Ilya Lavrenov [Sun, 12 Oct 2014 20:37:59 +0000 (13:37 -0700)]
cv::convertMaps
Ilya Lavrenov [Sun, 12 Oct 2014 20:13:24 +0000 (13:13 -0700)]
removed saturation from pyramids
Ilya Lavrenov [Sun, 12 Oct 2014 20:05:30 +0000 (13:05 -0700)]
fixes for cv::warpAffine
Ilya Lavrenov [Sun, 12 Oct 2014 19:40:41 +0000 (12:40 -0700)]
fixes in VResizeLinearVec_32s8u
Ilya Lavrenov [Sun, 12 Oct 2014 18:47:50 +0000 (11:47 -0700)]
cv::resize (INTER_LANCZOS4)
Ilya Lavrenov [Sun, 12 Oct 2014 17:45:46 +0000 (10:45 -0700)]
cv::resize (INTER_AREA CV_16S, CV_32F)
Ilya Lavrenov [Sun, 12 Oct 2014 17:16:59 +0000 (10:16 -0700)]
cv::compareHist (CV_COMP_INTERSECT)
Ilya Lavrenov [Sun, 12 Oct 2014 15:31:25 +0000 (08:31 -0700)]
cv::pow (integer power)
Ilya Lavrenov [Sun, 12 Oct 2014 11:54:27 +0000 (04:54 -0700)]
cv::polarToCart
Ilya Lavrenov [Sun, 12 Oct 2014 10:43:06 +0000 (03:43 -0700)]
cv::magnitude; cv::corner**
Ilya Lavrenov [Sun, 12 Oct 2014 10:04:03 +0000 (03:04 -0700)]
cv::phase; cv::cartToPolar
Ilya Lavrenov [Sat, 11 Oct 2014 10:38:42 +0000 (03:38 -0700)]
cv::pyrUp
Ilya Lavrenov [Sat, 11 Oct 2014 09:34:04 +0000 (02:34 -0700)]
cv::warpAffine
Ilya Lavrenov [Sat, 11 Oct 2014 08:51:20 +0000 (01:51 -0700)]
VResizeLinearVec_32s8u
Vadim Pisarevsky [Thu, 30 Oct 2014 10:45:02 +0000 (10:45 +0000)]
Merge pull request #3343 from mshabunin:doxygen-docs
Vadim Pisarevsky [Thu, 30 Oct 2014 10:44:21 +0000 (10:44 +0000)]
Merge pull request #3361 from mshabunin:fix-drawing
Alexander Alekhin [Wed, 29 Oct 2014 14:25:09 +0000 (14:25 +0000)]
Merge pull request #3318 from akarsakov:ocl_cvtcolor
Alexander Alekhin [Wed, 29 Oct 2014 13:53:56 +0000 (13:53 +0000)]
Merge pull request #3354 from vbystricky:oclopt_convertScaleAbs
Alexander Alekhin [Wed, 29 Oct 2014 11:18:06 +0000 (11:18 +0000)]
Merge pull request #3159 from ElenaGvozdeva:ocl_gemm
Alexander Alekhin [Tue, 28 Oct 2014 17:19:57 +0000 (17:19 +0000)]
Merge pull request #3379 from akarsakov:gftt_empty_src
Alexander Karsakov [Tue, 28 Oct 2014 13:06:36 +0000 (16:06 +0300)]
Don't process empty src image in GFTT
ElenaGvozdeva [Tue, 28 Oct 2014 12:18:31 +0000 (15:18 +0300)]
use LOCAL_SIZE+1
Alexander Karsakov [Tue, 28 Oct 2014 12:07:51 +0000 (15:07 +0300)]
Added optimized loading to YUV2RGB_422 kernel
Alexander Alekhin [Tue, 28 Oct 2014 10:23:18 +0000 (10:23 +0000)]
Merge pull request #3353 from Chuanbo-Weng:master
Alexander Alekhin [Tue, 28 Oct 2014 09:18:31 +0000 (09:18 +0000)]
Merge pull request #3359 from akarsakov:ocl_buildMaps
Jose-Luis Blanco-Claraco [Tue, 28 Oct 2014 08:46:18 +0000 (09:46 +0100)]
OpenCVConfig.cmake: Fix CMake CMP0012 warning
"TRUE"/"FALSE" should not be used as literals, as CMake has to decide
whether they are literals or variable names (sigh).
This patch should make it work with any version of CMake and with any
user-side decision on old vs. new policy behavior.
Alexander Karsakov [Mon, 27 Oct 2014 11:52:17 +0000 (14:52 +0300)]
Added loading 4 pixels in line instead of 2 to RGB[A] -> YUV(420) kernel
Alexander Karsakov [Mon, 27 Oct 2014 12:38:44 +0000 (15:38 +0300)]
Added checking that z is non zero to buildWarpPlaneMaps kernel
Alexander Karsakov [Fri, 24 Oct 2014 08:03:10 +0000 (11:03 +0300)]
Overload PlaneWarper::buildMaps method from base class
ElenaGvozdeva [Thu, 16 Oct 2014 07:24:44 +0000 (10:24 +0300)]
Some small fixes
Elena Gvozdeva [Thu, 4 Sep 2014 08:36:23 +0000 (12:36 +0400)]
use vectors
Elena Gvozdeva [Wed, 27 Aug 2014 10:58:01 +0000 (14:58 +0400)]
use local memory
Elena Gvozdeva [Fri, 22 Aug 2014 11:05:29 +0000 (15:05 +0400)]
naive implementation
Alexander Alekhin [Mon, 27 Oct 2014 10:03:29 +0000 (10:03 +0000)]
Merge pull request #3369 from vbystricky:fix_scaleAdd
Pat O'Keefe [Sun, 26 Oct 2014 01:11:02 +0000 (19:11 -0600)]
Add test for calib3d function decomposeProjectionMatrix(...)
Alexander Alekhin [Fri, 24 Oct 2014 16:01:21 +0000 (16:01 +0000)]
Merge pull request #3367 from akarsakov:ocl_image2d
Alexander Alekhin [Fri, 24 Oct 2014 16:00:20 +0000 (16:00 +0000)]
Merge pull request #3342 from alalek:perf_stability_check
vbystricky [Fri, 24 Oct 2014 10:05:39 +0000 (14:05 +0400)]
Move _dst.create() to the begining of scaleAdd function
Alexander Alekhin [Fri, 24 Oct 2014 12:59:13 +0000 (12:59 +0000)]
Merge pull request #3284 from vbystricky:oclopt_stereoBM
Alexander Alekhin [Fri, 24 Oct 2014 12:58:21 +0000 (12:58 +0000)]
Merge pull request #3368 from alalek:amd_clfinish
Alexander Karsakov [Fri, 24 Oct 2014 10:55:16 +0000 (13:55 +0300)]
Added extra checks to ocl::Image2D
Alexander Alekhin [Fri, 24 Oct 2014 10:29:38 +0000 (14:29 +0400)]
ocl: restore clFinish() in unmap() for AMD devices
This reverts commit
7d91b8efcd053304e5b0fb5afff06e151b2d4096.
Maksim Shabunin [Thu, 23 Oct 2014 14:43:24 +0000 (18:43 +0400)]
Disabled videoio_video_prop_fps test
This test does not return correct values with some software configurations.
Possible problem with ffmpeg on buildbot's linux-slave.
vbystricky [Mon, 29 Sep 2014 08:46:56 +0000 (12:46 +0400)]
Optimize OpenCL version of StereoBM function
Fix problems on NVidia devices.
Alexander Karsakov [Thu, 23 Oct 2014 11:23:37 +0000 (14:23 +0300)]
Added tests for Image2D
Maksim Shabunin [Thu, 23 Oct 2014 08:33:38 +0000 (12:33 +0400)]
Fixed imgcodecs tests
- Imgcodecs_Image, write_imageseq - assuming JPEG images have losses
- Imgcodecs_Drawing, c_regression - replaced assert calls
- Imgcodecs_Drawing - exact comparison with reference image
Miroslav Benes [Wed, 22 Oct 2014 14:54:39 +0000 (16:54 +0200)]
Add Triangle thresholding algorithm
Add Triangle method for automatic threshold computation next to the existing
Otsu's method. Triangle deals better with images whose histogram does not
contain dominant peak.
See paper Zack GW, Rogers WE, Latt SA.: Automatic measurement of sister
chromatid exchange frequency. J Histochem Cytochem. 1977 Jul;25(7):741-53.
Alexander Karsakov [Wed, 22 Oct 2014 14:43:29 +0000 (17:43 +0300)]
Small optimization for buildMaps kernels in case rowsPerWI > 1
Maksim Shabunin [Thu, 9 Oct 2014 10:04:00 +0000 (14:04 +0400)]
Basic doxygen documentation support
- updated existing Doxyfile.in
- added corresponding cmake instructions
- added some specific files (layout, icon)
- clean existing doxygen warnings
Vadim Pisarevsky [Wed, 22 Oct 2014 01:55:23 +0000 (01:55 +0000)]
Merge pull request #3352 from alucarded:BackgroundSubtractorMOG2_optimization
vbystricky [Mon, 20 Oct 2014 13:43:18 +0000 (17:43 +0400)]
Optimize OpenCL version of conversScaleAbs function
Alexander Karsakov [Tue, 14 Oct 2014 12:31:10 +0000 (16:31 +0400)]
Used direct float calculations
Alexander Karsakov [Mon, 6 Oct 2014 15:21:57 +0000 (19:21 +0400)]
Added OCL code for YUV422 -> RGB[A]|BGR[A] color conversion
Alexander Karsakov [Mon, 6 Oct 2014 15:19:44 +0000 (19:19 +0400)]
Added OCL code for RGB[A]|BGR[A] -> YUV_[YV12|IYUV] color conversion
Alexander Karsakov [Fri, 3 Oct 2014 08:29:15 +0000 (12:29 +0400)]
Added OCL code for YUV2GRAY_420 color conversion
Alexander Karsakov [Thu, 2 Oct 2014 14:10:14 +0000 (18:10 +0400)]
Added OCL code for YUV2BGR_YV12 and YUV2BGR_IYUV color conversions
Alexander Karsakov [Wed, 1 Oct 2014 14:08:07 +0000 (18:08 +0400)]
Added support for YUV2RGB[A]_NV21 and YUV2BGR[A]_NV21 conversion
Alexander Alekhin [Tue, 21 Oct 2014 11:23:19 +0000 (15:23 +0400)]
added performance validation check
Alexander Alekhin [Tue, 21 Oct 2014 11:23:06 +0000 (11:23 +0000)]
Merge pull request #3355 from ElenaGvozdeva:predictOptimalVectorWidth
ElenaGvozdeva [Tue, 21 Oct 2014 10:13:15 +0000 (13:13 +0300)]
Changed predictOptimalVectorWidth function, now it is possible to choose vector size.
Alexander Alekhin [Tue, 21 Oct 2014 08:59:23 +0000 (08:59 +0000)]
Merge pull request #3256 from vbystricky:oclopt_BFMatcher
Tomasz Posłuszny [Sun, 19 Oct 2014 21:50:37 +0000 (23:50 +0200)]
BackgroundSubtractorMOG2, learningRate=0 optimization
Chuanbo Weng [Mon, 20 Oct 2014 10:14:38 +0000 (18:14 +0800)]
Update hog.cpp
Update according to vbystricky's comments
Vadim Pisarevsky [Mon, 20 Oct 2014 06:50:34 +0000 (06:50 +0000)]
Merge pull request #3344 from mshabunin:java_automation
Vadim Pisarevsky [Mon, 20 Oct 2014 06:49:18 +0000 (06:49 +0000)]
Merge pull request #3349 from a-wi:DShow_Strsafe_header_v3
Vadim Pisarevsky [Mon, 20 Oct 2014 06:41:51 +0000 (06:41 +0000)]
Merge pull request #3292 from mshabunin:fix-ios-warnings
Chuanbo Weng [Mon, 20 Oct 2014 03:50:46 +0000 (11:50 +0800)]
Correctly enable OpenCL mode in tapi's hog example.
For current OpenCV-CL architecture, if the data buffer
allocated in UMat are cpu buffer(not ocl buffer) under
cpu mode, and then pass this UMat to an OpenCL kernel
as an argument, the OpenCL path will fail and fallback
to cpu mode. Take HOGDescriptor::oclSvmDetector as an example:
ocl::setUseOpenCL(false);
//data allocated in hog.oclSvmDetector will be cpu buffer
hog.setSVMDetector(HOGDescriptor::getDaimlerPeopleDetector());
ocl::setUseOpenCL(true);
//We enabled OpenCL, but hog.oclSvmDetector are cpu buffer,
//so it will fail in the function ocl_classify_hists
//when reach to this line
//idx = k.set(idx, ocl::KernelArg::PtrReadOnly(detector));
hog.detectMultiScale(img, found, hit_threshold, win_stride,
Size(0, 0), scale, gr_threshold);
Similar problems heppen on img_aux and img. So we should re-define
or re-set these UMat when do mode switch (CPU -> OpenCL) in order
to make their data be allocated by ocl and then OpenCL path will
succeed.
Signed-off-by: Chuanbo Weng <chuanbo.weng@intel.com>
Vadim Pisarevsky [Sat, 18 Oct 2014 18:52:35 +0000 (18:52 +0000)]
Merge pull request #3348 from vpisarev:refactor_algorithms2
Artur Wieczorek [Sat, 18 Oct 2014 18:27:49 +0000 (20:27 +0200)]
Include strsafe.h to compile successfully under MinGW in Debug mode.
Vadim Pisarevsky [Sat, 18 Oct 2014 17:58:34 +0000 (21:58 +0400)]
fixed MSVC warnings and iOS framework compile error
Vadim Pisarevsky [Sat, 18 Oct 2014 17:00:31 +0000 (17:00 +0000)]
Merge pull request #3345 from D-Alex:master
Vadim Pisarevsky [Sat, 18 Oct 2014 16:57:35 +0000 (16:57 +0000)]
Merge pull request #3347 from Nerei:memory_leak_in_viz_opencv30
Vadim Pisarevsky [Sat, 18 Oct 2014 16:53:39 +0000 (20:53 +0400)]
fixed bug in automatic SVM training
Vadim Pisarevsky [Sat, 18 Oct 2014 16:44:26 +0000 (20:44 +0400)]
quickly corrected the previous refactoring of features2d: moved from set(SOME_PROP, val) to setSomeProp(val)
Anatoly Baksheev [Sat, 18 Oct 2014 14:17:42 +0000 (18:17 +0400)]
viz: fixed memory leak
Alexander Duda [Sat, 18 Oct 2014 12:22:27 +0000 (14:22 +0200)]
fix warning in template Scalar::mul gcc < 4.8
In this particular case t shadows transpose of the base class Matx:
types.hpp:1805:14: warning: declaration of ‘t’ shadows a member of
'this' [-Wshadow]
Changelog gcc 4.8: The option -Wshadow no longer warns if a declaration
shadows a function declaration.
This warning is problematic because it prevents the module
opencv_contrib/modules/ruby to pass the build process
vbystricky [Tue, 23 Sep 2014 11:13:46 +0000 (15:13 +0400)]
Optimize OpenCL version of BFMatcher
Maksim Shabunin [Fri, 17 Oct 2014 15:36:24 +0000 (19:36 +0400)]
Refactoring in gen_java.py script