Vladislav Vinogradov
bff0fad6c3
gpu TVL1 Optical Flow optimization:
...
do not calculate sum of error in every round of iteration;
instead the error will be summed every 2nd times or more,
if the previous sum of error is too far away from threshold.
2013-08-27 11:21:41 +04:00
Vladislav Vinogradov
21c4753fed
used global memory access for up-scaling
2013-08-23 18:08:06 +04:00
Vladislav Vinogradov
d1f6a23abf
improved texture usage:
...
* use explicit extrapolation only for sub-matrixes
* use built-in interpolation for INTER_NEAREST mode
2013-08-23 13:57:12 +04:00
Vladislav Vinogradov
241cc417f9
added custom implementation for nearest and linear interpolation
2013-08-23 13:57:12 +04:00
Vladislav Vinogradov
3b05acf936
reorganize code for further modifiction
2013-08-23 13:56:35 +04:00
Vladislav Vinogradov
f826bd8bce
removed NPP implementation
2013-08-23 12:16:41 +04:00
Roman Donchenko
fe4c7ae4ae
Merge pull request #1317 from jet47:gpu-tvl1-optimization
2013-08-22 11:55:24 +04:00
Roman Donchenko
e9a28f66ee
Normalized file endings.
2013-08-21 18:59:25 +04:00
Roman Donchenko
f55740da70
Deleted all trailing whitespace.
2013-08-21 18:59:24 +04:00
Vladislav Vinogradov
de214950c4
minor gpu TVL1 optical flow optimization:
...
don't calc diff term if it is not used for epsilon criterion
2013-08-21 18:17:45 +04:00
Roman Donchenko
1a7c2eefb8
Merge pull request #1306 from jet47:gpu-bitwize-fix
2013-08-21 16:18:31 +04:00
Roman Donchenko
84a9755597
Merge pull request #1215 from jet47:cuda-5.5-sanity
2013-08-21 13:25:28 +04:00
Vladislav Vinogradov
cb2985c526
fixed gpu bitwize operation:
...
incorrect types for template instantiation
2013-08-20 16:28:45 +04:00
Vladislav Vinogradov
cfbd791465
fixed gpu cvtColor for BGR <-> YUV
2013-08-19 11:56:58 +04:00
kamjagin
c24c1a5926
added HAVE_QTKIT to tests
2013-08-14 13:40:50 +02:00
Roman Donchenko
d9f71b762f
Deleted almost all of the precomp.cpp files.
...
Looks like we don't actually use them; when creating precompiled headers
with Visual C++, a dummy .cpp is created at build time.
2013-08-08 12:43:07 +04:00
StevenPuttemans
ed76b2f98f
Like mentioned by Andrei Pavlenko after merging pullrequest #1206 , a wrong
...
ReST directive was used. Also fixed some other ReST directives that were
not correct and removed some warnings during buildbot checks.
2013-08-07 14:49:33 +02:00
Andrey Pavlenko
ee8687ba19
Merge pull request #1214 from StevenPuttemans:feature_3169_others
2013-08-02 17:59:06 +04:00
Vladislav Vinogradov
d13ecd0d76
increase epsilon for GEMM & Convolve sanity tests for CUDA 5.5
2013-08-02 17:31:34 +04:00
StevenPuttemans
f28496c628
Added sample links in documentation for python2, ocl and gpu
2013-08-02 14:05:08 +02:00
Vladislav Vinogradov
55cb26551f
use NPP version for Fermi
2013-08-02 09:05:21 +04:00
Vladislav Vinogradov
a28cb99e88
optimized version of histEven for CV_8UC1
2013-08-01 18:00:05 +04:00
Vladislav Vinogradov
bfca1afd9e
rewrite histEven test
2013-08-01 18:00:05 +04:00
Vladislav Vinogradov
c8d1fccdac
compare cv::gpu::calcHist with cv::calcHist
2013-08-01 18:00:05 +04:00
Andrey Pavlenko
6a40067424
Merge pull request #1206 from StevenPuttemans:feature_3169
2013-07-31 14:31:21 +04:00
StevenPuttemans
0c5244103f
Added all *.cpp samples reference to the 2.4.6 documentation as discussed in feature #3169
...
- Used lunix style slashes - works on most systems
- Removed all trailing whitespaces
2013-07-31 11:20:05 +02:00
Andrey Pavlenko
f7ce70f70a
Merge pull request #1110 from SpecLad:config-cleanup
2013-07-30 17:39:35 +04:00
Vladislav Vinogradov
dfd2957841
fixed gpu_perf4au perf test (added cuda impl)
2013-07-26 11:13:39 +04:00
Vladislav Vinogradov
ebe7ff991f
fixed compilation without CUBLAS and CUFFT
2013-07-18 12:04:33 +04:00
Roman Donchenko
ac39bfb4cc
Remove HAVE_CVCONFIG_H - it's always defined.
2013-07-10 16:02:07 +04:00
Roman Donchenko
156155addb
Merge pull request #1018 from jet47:cuda-5.5-arm
2013-06-28 14:26:45 +04:00
Roman Donchenko
c244f000ab
Merge pull request #1030 from SpecLad:ts-impl-variants
2013-06-24 18:28:13 +04:00
Roman Donchenko
3ea4836a0a
Changed the impls argument to be an array name.
...
Turns out, you can't use preprocessor directives inside macro arguments.
Who'd have thought?
2013-06-20 15:18:35 +04:00
Roman Donchenko
51a672ec40
Disabled the cuda variant when CUDA is not available.
2013-06-19 19:16:18 +04:00
Roman Donchenko
936236e4b1
Extended the CPU/GPU selection mechanism in performance tests.
...
Now it allows choosing between arbitrary implementation variants.
2013-06-19 17:55:42 +04:00
Vladislav Vinogradov
371a9cd833
fixed build with CUDA 5.5 on arm platforms
2013-06-18 17:46:57 +04:00
Vladislav Vinogradov
1492b20472
fix gpu warnings with signed/unsigned char
2013-06-18 17:44:49 +04:00
Vladislav Vinogradov
0367a7f992
link with nvcuvenc and ffmpeg libraries only if WITH_NVCUVID is enabled
2013-06-13 16:46:34 +04:00
Roman Donchenko
7f9675bcc0
Merge pull request #953 from jet47:fix-bug-3069
2013-06-05 17:56:35 +04:00
Roman Donchenko
087db2949a
Merge pull request #948 from jet47:cuda-5.5-support
2013-06-05 17:04:22 +04:00
Vladislav Vinogradov
31a5f7ef3c
fixed bug #3069 (infinite loop in GPU LBP Cascade detectMultiScale)
2013-06-05 14:08:55 +04:00
Vladislav Vinogradov
89f3c40d79
fixed BroxOpticalFlow sanity test (increase epsilon value)
...
+ interpolateFrames and createOpticalFlowNeedleMap
2013-06-04 15:01:06 +04:00
Vladislav Vinogradov
4a770535c4
fixed BoxFilter sanity test (different rounding results)
2013-06-04 14:59:47 +04:00
Vladislav Vinogradov
516e5b2563
fixed BroxOpticalFlow regression test
...
the output of BroxOpticalFlow differs a bit in CUDA 5.5
2013-06-04 13:58:45 +04:00
Vladislav Vinogradov
918381875a
rewrite gpu/device/vec_math.hpp file
...
old version isn't compiled with CUDA 5.5
new version doesn't depend on functional.hpp
2013-06-04 13:57:35 +04:00
Vladislav Vinogradov
3aea7e8f8d
fixed gpu module build on arm platform
...
links with CUDA driver library only if we use video encoding/decoding
2013-06-04 12:51:36 +04:00
Roman Donchenko
75cf5cc4ee
Merge pull request #943 from jet47:cuda-5.5-support
2013-06-03 16:08:23 +04:00
Vladislav Vinogradov
bcf8bdb401
fixed constructors for functional objects (added __host__ modifier)
2013-06-03 14:41:23 +04:00
Vladislav Vinogradov
58e472754a
fixed norm diff function (it uses pre-allocated buffer now)
2013-06-03 13:37:44 +04:00
Vladislav Vinogradov
0521e8908c
fixed NPP error constants usage
2013-06-03 13:37:36 +04:00
