0172bd6beb
Merge pull request #3756 from eliteraspberries:cast-double
2015-03-02 10:52:28 +00:00
e02d6e2c8f
Merge pull request #3766 from StevenPuttemans:add_GPU_old_models
2015-03-02 10:49:38 +00:00
2c0ef660bf
Add GPU models in seperate folder, only old models are supported and they were replaced to make the CPU interface work.
...
This solves the undesired consequence of breaking the GPU interface.
2015-02-27 14:15:30 +01:00
21bbed7baf
fix gpu HOG implementation:
...
move hog::set_up_constants from constructor to compute method
if user changed CUDA device between constructor and computation,
some variables were uninitialized
2015-02-27 14:28:12 +03:00
c147ab1e85
set epsilon for gpu OpticalFlowBM to 1e-6, since it uses floating point arithmetic
2015-02-25 13:16:36 +03:00
b7be49426d
Cast some image coordinates and sizes to double.
2015-02-24 12:21:34 -05:00
7bb8c50080
disable gpu::matchTemplate tests
2015-02-04 11:21:17 +03:00
c0a8d4e1ef
Fix of HaarCascade GPU test for aarch64.
2014-12-28 11:49:25 +03:00
d00f36ec75
fix tests for gpu::matchTemplate:
...
use ASSERT_FLOAT_EQ to compare float values, it is more robust for
large values
2014-12-25 17:43:13 +03:00
73d1ea9a52
fix bug 2432 for 2.4 branch
2014-12-15 11:56:42 +01:00
eeb997261d
get rid of cuda.h usage
2014-11-25 18:29:56 +03:00
da6a38f9f6
NCV.cu: include <algorithm>
...
It's needed for std::max.
2014-11-13 16:16:26 +03:00
529af7a90e
Merge pull request #3364 from mshabunin:doxygen-docs-24
2014-10-30 10:42:25 +00:00
84f33d0578
minor fix for StereoCSBP data cost compute kernel and test
2014-10-23 19:33:44 +04:00
5c07e0b6d3
increase epsilons for some tests, which functions use floating point arithm
2014-10-23 19:33:42 +04:00
e0827069c1
avoid pointer arithmetic on register memory in color conversion
2014-10-23 18:39:23 +04:00
9b55c09999
Basic doxygen documentation support
...
- updated existing Doxyfile.in
- added corresponding cmake instructions
- added some specific files (layout, icon)
- clean existing doxygen warnings
Conflicts:
CMakeLists.txt
doc/CMakeLists.txt
modules/core/include/opencv2/core.hpp
modules/core/include/opencv2/core/base.hpp
modules/core/include/opencv2/core/cuda.inl.hpp
modules/core/include/opencv2/core/mat.hpp
modules/core/include/opencv2/core/matx.hpp
modules/core/include/opencv2/core/types.hpp
modules/flann/include/opencv2/flann/lsh_table.h
modules/imgproc/include/opencv2/imgproc.hpp
2014-10-23 18:27:26 +04:00
eaaa2d27d5
fix CUDA LUT implementation
...
In CUDA 6.0 there was a bug in NPP LUT implementation (invalid results when
src == 255). In CUDA 6.5 the bug was fixed.
Replaced NPP LUT call with own implementation (ported from master branch)
to be independant from CUDA Toolkit version.
2014-08-28 14:47:26 +04:00
6eb26c1519
Merge pull request #3109 from jet47:gpu-test-fixes
2014-08-22 07:22:23 +00:00
ebe36d6e7c
fix CUDA cvtColor after corresponding change in CPU version
...
see https://github.com/Itseez/opencv/pull/3137
2014-08-21 14:14:06 +04:00
d558260a8e
fixing GT and GE comparison symbol
2014-08-18 13:33:40 +02:00
62f27b28ed
fix BGR->BGR5x5 color convertion
2014-08-15 14:10:15 +04:00
599f5ef51b
use downscaled frames in FGDStatModel test
2014-08-15 13:42:25 +04:00
86e12b6074
increase epsilon for ResizeSameAsHost test
2014-08-15 13:42:06 +04:00
5dff283b39
increase epsilon for TVL1 sanity test
2014-08-15 13:41:47 +04:00
da9be8231f
fix cv::gpu::resize for INTER_LINEAR, now it produces the same result as CPU version
2014-08-01 11:33:29 +04:00
60a5ada454
Build fixes for CUDA 6.5
2014-07-14 14:26:50 +04:00
ecec53f509
fixed docs
2014-07-09 17:48:28 +04:00
1f72873c55
fix cv::gpu::resize function
...
add missing stream parameter to call_resize_linear_glob
2014-05-14 10:31:28 +04:00
a01659825c
remove unused variable in loadFromXML()
2014-04-09 22:53:36 +03:00
d822d6f101
disable GeneralizedHough tests on gcc 4.7
2014-04-07 09:49:17 +04:00
f8d922ad3e
disable gpu::GeneralizedHough compilation on gcc 4.7
2014-04-04 10:31:33 +04:00
0d5ffec521
Merge pull request #2553 from ilya-lavrenov:typo
2014-03-31 23:21:07 +04:00
9b9769169f
doc typos
2014-03-31 19:57:05 +04:00
6d368dc1e7
increase epsilon for MOG2::getBackgroundImage test:
...
it uses float arithmetics
2014-03-30 18:02:03 -07:00
0a16d93e1d
Fixed an issue with weight assignment causing the resulting GMM weights to be unsorted in the CUDA and OCL versions of BackgroundSubtractorMOG2
2014-03-18 17:26:24 +02:00
099ea91823
typos
2014-03-04 12:47:40 +04:00
af8aa8a2fa
Merge pull request #2433 from ilya-lavrenov:typos
2014-03-03 12:20:35 +04:00
eedf86402d
typos
2014-03-02 21:04:17 +04:00
7fd21d217a
fix bug #3562 :
...
add missing __syncthreads to edgesHysteresisLocalKernel
2014-02-20 10:24:52 +04:00
da9bdf9c15
fixes typo and unused variables
2014-02-18 14:16:54 +04:00
150e522bee
fix bug #3552 :
...
replace std::swap with own code
2014-02-16 12:17:17 +04:00
1ce5165cb7
temporary disable performance test for alphaComp function
2014-02-14 14:27:43 +04:00
eb247d826f
temporary disable perf test for StereoBeliefPropagation
2014-02-13 17:25:59 +04:00
9e69e2a07a
increase epsilon for AlphaComp sanity test for integer input
2014-02-13 17:16:43 +04:00
8b44a42a40
decrease input size for several gpu tests to fix "timed out" error:
...
* BruteForceNonLocalMeans
* OpticalFlowBM
2014-02-10 11:50:14 +04:00
1d68a89947
Merge pull request #2303 from jet47:gpu-canny-fix
2014-02-07 19:03:14 +04:00
5dbdadb769
fixed several bugs in CUDA Canny implementation:
...
* out of border access in edgesHysteresisLocalKernel
* incorrect usage of atomicAdd
2014-02-07 16:04:29 +04:00
bfc27271e2
Revert "disable gpu Canny and HoughCircles perf tests:"
...
This reverts commit dbce90692a
2014-02-07 14:55:06 +04:00
5d099df578
Revert "disable CUDA generalized Hough Transform"
...
This reverts commit 33d42b740c
2014-02-06 18:02:14 +04:00
Vadim Pisarevsky