Ilya Lavrenov [Mon, 29 Dec 2014 12:38:49 +0000 (15:38 +0300)]
SSE2 optimization of cv::integral
Vadim Pisarevsky [Wed, 24 Dec 2014 11:20:27 +0000 (11:20 +0000)]
Merge pull request #3523 from jet47:fix-cuda-buffer-pool
Vadim Pisarevsky [Wed, 24 Dec 2014 11:20:06 +0000 (11:20 +0000)]
Merge pull request #3527 from jet47:cuda-deprivate-old-device-layer
Vadim Pisarevsky [Wed, 24 Dec 2014 11:14:54 +0000 (11:14 +0000)]
Merge pull request #3541 from jet47:find-cuda-update
Vadim Pisarevsky [Wed, 24 Dec 2014 09:02:23 +0000 (09:02 +0000)]
Merge pull request #3532 from oresths:filter_neon
Vadim Pisarevsky [Wed, 24 Dec 2014 09:00:32 +0000 (09:00 +0000)]
Merge pull request #3538 from alalek:icv_fix_package
Vladislav Vinogradov [Tue, 23 Dec 2014 14:48:18 +0000 (17:48 +0300)]
[FindCUDA] improvements for cross-platform build
* improve `CUDA_TARGET_CPU_ARCH` cache initialization,
allow to override initial value from calling script;
* add `CUDA_TARGET_OS_VARIANT` option to select OS variant;
* add `CUDA_TARGET_TRIPLET` option to select target triplet from
`${CUDA_TOOLKIT_ROOT_DIR}/targets` folder;
* remove `CUDA_TOOLKIT_TARGET_DIR` option, now it is calculated from
`CUDA_TARGET_TRIPLET`, the old approach still can be used for compatibility;
* for CUDA 6.5 and newer try to locate static libraries too, because
in 6.5 toolkit for ARM cross compilation only static libraries are included.
Vladislav Vinogradov [Tue, 23 Dec 2014 14:47:04 +0000 (17:47 +0300)]
update FindCUDA CMake module to the latest version from upstream
Vladislav Vinogradov [Thu, 18 Dec 2014 14:56:31 +0000 (17:56 +0300)]
minor reorganization for CUDA doxygen groups:
move main CUDA group to modules/core/cuda.hpp
Vladislav Vinogradov [Thu, 18 Dec 2014 14:55:48 +0000 (17:55 +0300)]
mark old CUDA device layer as deprecated and remove it from doxygen documentation
add a note to use new cudev module as a replacement
Vladislav Vinogradov [Mon, 22 Dec 2014 08:33:39 +0000 (11:33 +0300)]
fix null stream initialization for multi-gpu systems
Vladislav Vinogradov [Fri, 19 Dec 2014 16:33:32 +0000 (19:33 +0300)]
move StackAllocator to cpp file
it is internal class, no need to export it
Vladislav Vinogradov [Wed, 17 Dec 2014 15:51:15 +0000 (18:51 +0300)]
fix cuda::BufferPool deinitialization
The deinitialization of BufferPool internal objects is controled by global
object, but it depends on other global objects, which leads to errors
caused by undefined deinitialization order of global objects.
I merge global objects initialization into single class, which performs
initialization and deinitialization in correct order.
Alexander Alekhin [Tue, 23 Dec 2014 13:10:47 +0000 (16:10 +0300)]
IPPICV: don't use full paths in dependencies
Vadim Pisarevsky [Tue, 23 Dec 2014 13:38:23 +0000 (13:38 +0000)]
Merge pull request #3529 from jet47:fix-linux-install
Vadim Pisarevsky [Tue, 23 Dec 2014 13:36:40 +0000 (13:36 +0000)]
Merge pull request #3536 from mshabunin:doxygen-intro
Maksim Shabunin [Mon, 22 Dec 2014 14:21:37 +0000 (17:21 +0300)]
Fixed some mistakes
Maksim Shabunin [Tue, 9 Dec 2014 15:45:36 +0000 (18:45 +0300)]
Tutorial: documenting OpenCV
Vadim Pisarevsky [Mon, 22 Dec 2014 11:58:01 +0000 (11:58 +0000)]
Merge pull request #3513 from mshabunin:compat-30
Vadim Pisarevsky [Mon, 22 Dec 2014 11:14:29 +0000 (11:14 +0000)]
Merge pull request #3520 from JoeHowse:master
Vadim Pisarevsky [Mon, 22 Dec 2014 10:58:07 +0000 (10:58 +0000)]
Merge pull request #3524 from jet47:fix-cuda-warnings
Vadim Pisarevsky [Mon, 22 Dec 2014 10:57:06 +0000 (10:57 +0000)]
Merge pull request #3525 from jet47:fix-cudev-tests
Vadim Pisarevsky [Mon, 22 Dec 2014 10:44:47 +0000 (10:44 +0000)]
Merge pull request #3490 from oresths:symmcolumnsmall_fix
Vadim Pisarevsky [Mon, 22 Dec 2014 10:39:42 +0000 (10:39 +0000)]
Merge pull request #3512 from vins31:OpenNi2_AsusXtion
Vadim Pisarevsky [Mon, 22 Dec 2014 10:29:31 +0000 (10:29 +0000)]
Merge pull request #3499 from StevenPuttemans:fix_2432
Vadim Pisarevsky [Mon, 22 Dec 2014 10:27:33 +0000 (10:27 +0000)]
Merge pull request #3510 from boaz001:feature-4057
Vadim Pisarevsky [Mon, 22 Dec 2014 10:26:30 +0000 (10:26 +0000)]
Merge pull request #3516 from ana-GT:openni2_defaultMode
Vadim Pisarevsky [Mon, 22 Dec 2014 10:25:47 +0000 (10:25 +0000)]
Merge pull request #3518 from wangyan42164:ocl_cascade_detect
Vadim Pisarevsky [Mon, 22 Dec 2014 10:25:00 +0000 (10:25 +0000)]
Merge pull request #3519 from fvgoto:patch-1
Vadim Pisarevsky [Mon, 22 Dec 2014 10:22:19 +0000 (10:22 +0000)]
Merge pull request #3528 from ilya-lavrenov:update_android_cmake
Vladislav Vinogradov [Thu, 18 Dec 2014 08:37:26 +0000 (11:37 +0300)]
increase epsilons for tests due to different optimizations (IPP vs CUDA, float vs double)
Vladislav Vinogradov [Thu, 18 Dec 2014 08:36:43 +0000 (11:36 +0300)]
update cudev color conversions according to the latest changes in CPU code
Vladislav Vinogradov [Thu, 18 Dec 2014 08:35:45 +0000 (11:35 +0300)]
disable -Wshadow warning for CUDA modules:
it is generated by CUDA headers and we can't fix it
orestis [Sun, 21 Dec 2014 19:27:03 +0000 (21:27 +0200)]
Change DescriptorExtractor_ORB regression test
to compensate for neon ieee754 non-compliancy.
Also changed the comparison between max valid and calculated distance to
make the error message more accurate (in case curMaxDist == maxDist)
orestis [Sat, 20 Dec 2014 15:14:21 +0000 (17:14 +0200)]
Change gaussianBlur5x5 perf test epsilon
Set it 1 instead of 0.001, as is already done in gaussianBlur3x3. That
will allow integer destination matrices that are not exactly the same,
but very close to the expected result, to pass the test.
orestis [Fri, 19 Dec 2014 20:51:42 +0000 (22:51 +0200)]
SymmRowSmallVec_32f 1x5 asymm
NEON speedup: 2.31x
Auto-vect speedup: 2.26x
Test kernel: [-0.9432, -1.1528, 0, 1.1528, 0.9432]
orestis [Fri, 19 Dec 2014 20:47:06 +0000 (22:47 +0200)]
SymmRowSmallVec_32f 1x5
NEON speedup: 2.36x
Auto-vect speedup: 2.36x
Test kernel: [0.1, 0.2408, 0.3184, 0.2408, 0.1]
orestis [Fri, 19 Dec 2014 20:44:39 +0000 (22:44 +0200)]
SymmColumnVec_32f16s asymm
NEON speedup: 9.46x
Auto-vect speedup: 1x
Test kernel: [-0.9432, -1.1528, 0, 1.1528, 0.9432]
orestis [Fri, 19 Dec 2014 20:42:31 +0000 (22:42 +0200)]
SymmColumnVec_32f16s
NEON speedup: 8.64x
Auto-vect speedup: 1x
Test kernel: [0.1, 0.2408, 0.3184, 0.2408, 0.1]
orestis [Fri, 19 Dec 2014 20:40:55 +0000 (22:40 +0200)]
SymmColumnSmallVec_32s16s 3x1 asymm
NEON speedup: 2.12x
Auto-vect speedup: 1.01x
Test kernel: [-2, 0, 2]
orestis [Fri, 19 Dec 2014 20:37:52 +0000 (22:37 +0200)]
SymmColumnSmallVec_32s16s [-1, 0, 1]
NEON speedup: 3.27x
Auto-vect speedup: 1.01x
orestis [Fri, 19 Dec 2014 20:36:46 +0000 (22:36 +0200)]
SymmColumnSmallVec_32s16s 3x1
NEON speedup: 1.75x
Auto-vect speedup: 1x
orestis [Fri, 19 Dec 2014 20:35:52 +0000 (22:35 +0200)]
SymmColumnSmallVec_32s16s [3, 10, 3] Scharr
NEON speedup: 2.04x
Auto-vect speedup: 1x
orestis [Fri, 19 Dec 2014 20:34:11 +0000 (22:34 +0200)]
SymmColumnSmallVec_32s16s [1, -2, 1]
NEON speedup: 2.75x
Auto-vect speedup: 1.01x
orestis [Fri, 19 Dec 2014 20:33:11 +0000 (22:33 +0200)]
SymmColumnSmallVec_32s16s [1, 2, 1]
NEON speedup: 2.66x
Auto-vect speedup: 1x
orestis [Fri, 19 Dec 2014 20:29:54 +0000 (22:29 +0200)]
SymmColumnVec_32s8u asymm
NEON speedup: 2.95x
Auto-vect speedup: 1x
Test kernel: [-0.9432, -1.1528, 0, 1.1528, 0.9432]
orestis [Fri, 19 Dec 2014 20:26:41 +0000 (22:26 +0200)]
SymmColumnVec_32s8u
NEON speedup: 1.96x
Auto-vect speedup: 1x
Test kernel: [0.0708, 0.2445, 0.3694, 0.2445, 0.0708]
orestis [Fri, 19 Dec 2014 20:23:09 +0000 (22:23 +0200)]
SymmRowSmallVec_8u32s 1x5 asymm
NEON speedup: 3.14x
Auto-vect speedup: 1.6x
Test kernel: [-5, -2, 0, 2, 5]
orestis [Fri, 19 Dec 2014 20:15:37 +0000 (22:15 +0200)]
SymmRowSmallVec_8u32s 1x3 asymmetric
NEON speedup: 1.95x
Auto-vect speedup: 1.17x
Test kernel: [-2, 0, 2]
orestis [Fri, 19 Dec 2014 20:11:52 +0000 (22:11 +0200)]
SymmRowSmallVec_8u32s [-1, 0, 1]
NEON speedup: 1.84x
Auto-vect speedup: 1.2x
orestis [Fri, 19 Dec 2014 20:10:58 +0000 (22:10 +0200)]
SymmRowSmallVec_8u32s 1x5 general
NEON speedup: 3.86x
Auto-vect speedup: 1.67x
Test kernel: [0.0708, 0.2445, 0.3694, 0.2445, 0.0708]
orestis [Fri, 19 Dec 2014 20:08:23 +0000 (22:08 +0200)]
SymmRowSmallVec_8u32s 1x3 general
NEON speedup: 2.56x
Auto-vect speedup: 1.26x
Test kernel: [1, 3, 1]
orestis [Fri, 19 Dec 2014 20:02:57 +0000 (22:02 +0200)]
SymmRowSmallVec_8u32s [1, 2, 1]
NEON speedup: 4.1
Auto-vect speedup: 2.25
Maksim Shabunin [Thu, 18 Dec 2014 14:56:24 +0000 (17:56 +0300)]
Included c-headers for better 2.4 compatibility
Vladislav Vinogradov [Fri, 12 Dec 2014 15:42:15 +0000 (18:42 +0300)]
remove opencv_testing.sh installation to /etc/profile.d/
opencv_run_all_tests_unix.sh already exports OPENCV_TEST_DATA_PATH
Ilya Lavrenov [Thu, 18 Dec 2014 16:31:10 +0000 (19:31 +0300)]
update android.toolchain.cmake
Vadim Pisarevsky [Thu, 18 Dec 2014 14:50:13 +0000 (14:50 +0000)]
Merge pull request #3517 from wangyan42164:ocl_denoising_2
Joe Howse [Wed, 17 Dec 2014 01:37:14 +0000 (21:37 -0400)]
For Java compatibility, avoiding explicit pseudonym in enum
Joe Howse [Wed, 17 Dec 2014 00:07:37 +0000 (20:07 -0400)]
Allow BGR, RGB, or gray capture via libv4l
Florian Verdet [Tue, 16 Dec 2014 11:46:07 +0000 (12:46 +0100)]
fix obvious copy+paste typo in computeDistance()
(with this else-branch, argument contour2 would not be used at all)
Yan Wang [Tue, 16 Dec 2014 08:21:05 +0000 (16:21 +0800)]
Use preprocessor for constant values in OpenCL kernel instead of
the parameter variable.
It could improve the performance of
OCL_Cascade_Image_MinSize_CascadeClassifier.CascadeClassifier/*.
Especially,
OCL_Cascade_Image_MinSize_CascadeClassifier.CascadeClassifier/15
OCL_Cascade_Image_MinSize_CascadeClassifier.CascadeClassifier/16
could be improved about 2% in Intel platform.
Signed-off-by: Yan Wang <yan.wang@linux.intel.com>
Yan Wang [Tue, 16 Dec 2014 06:25:24 +0000 (14:25 +0800)]
Remove unnecessary local variable "size".
It should also be better for optimizing when compiling kernel.
Signed-off-by: Yan Wang <yan.wang@linux.intel.com>
Ana Huaman Quispe [Mon, 15 Dec 2014 23:16:25 +0000 (18:16 -0500)]
Changed CV_DbgAssert to CV_Assert when setting video modes for OpenNI2. Otherwise, in release mode the default modes never get set
Vadim Pisarevsky [Mon, 15 Dec 2014 18:07:09 +0000 (18:07 +0000)]
Merge pull request #3491 from mshabunin:ios-extra-world
Vadim Pisarevsky [Mon, 15 Dec 2014 18:04:01 +0000 (18:04 +0000)]
Merge pull request #3473 from TobyWanKenobi:master
Vadim Pisarevsky [Mon, 15 Dec 2014 18:03:01 +0000 (18:03 +0000)]
Merge pull request #3448 from Ashod:highgui_trackbar_improvements
StevenPuttemans [Wed, 10 Dec 2014 12:41:22 +0000 (13:41 +0100)]
fix documentation bug 2432
vincentweb [Sun, 14 Dec 2014 19:52:24 +0000 (20:52 +0100)]
Added better OpenNI2 support to the Asus Xtion and Occipital Structure
sensors which do not have image generators.
Boaz Stolk [Fri, 12 Dec 2014 13:21:47 +0000 (14:21 +0100)]
add option to link to dynamic IPP libraries
Vadim Pisarevsky [Tue, 9 Dec 2014 14:12:22 +0000 (14:12 +0000)]
Merge pull request #3495 from taka-no-me:fix_android_arm64
Vadim Pisarevsky [Tue, 9 Dec 2014 14:06:23 +0000 (14:06 +0000)]
Merge pull request #3479 from mshabunin:doxygen-tutorials
Vadim Pisarevsky [Tue, 9 Dec 2014 14:05:35 +0000 (14:05 +0000)]
Merge pull request #3492 from Dmitry-Me:reduceVariablesScope1
Andrey Kamaev [Tue, 9 Dec 2014 09:06:56 +0000 (12:06 +0300)]
Fix build for Android arm64-v8a
Alexander Alekhin [Mon, 8 Dec 2014 15:49:21 +0000 (15:49 +0000)]
Merge pull request #3486 from Chuanbo-Weng:repeat-perf
Dmitry-Me [Mon, 8 Dec 2014 15:21:17 +0000 (18:21 +0300)]
Reduce variables scope, move declaration to first use point
TobyWanKenobi [Mon, 8 Dec 2014 08:21:55 +0000 (09:21 +0100)]
Removed "FORMAT_XML" to generalize format
Removed the "FileStorage::FORMAT_XML" to generalize format of the String
model.
orestis [Sat, 6 Dec 2014 17:53:07 +0000 (19:53 +0200)]
Fix SymmColumnSmallVec_32s16s SSE code
Specifically, in general antisymmetric case src[-1] (S0) should be
subtracted from from src[1] (S2), and not the opposite.
Ashod Nakashian [Sat, 27 Sep 2014 21:27:04 +0000 (17:27 -0400)]
highgui: Support to change trackbar count in setTrackbarPos and replaced deprecated CreateToolbarEx in Windows.
Vadim Pisarevsky [Sat, 6 Dec 2014 10:04:22 +0000 (10:04 +0000)]
Merge pull request #3487 from taka-no-me:ctrl_s_gtk
Vadim Pisarevsky [Sat, 6 Dec 2014 08:09:01 +0000 (08:09 +0000)]
Merge pull request #3472 from Dmitry-Me:preventDivByZero
Vadim Pisarevsky [Sat, 6 Dec 2014 08:08:44 +0000 (08:08 +0000)]
Merge pull request #3475 from fhsm:master
Vadim Pisarevsky [Sat, 6 Dec 2014 08:08:27 +0000 (08:08 +0000)]
Merge pull request #3478 from akarsakov:rect_fix_typo
Vadim Pisarevsky [Sat, 6 Dec 2014 08:07:39 +0000 (08:07 +0000)]
Merge pull request #3481 from mshabunin:python-test
Maksim Shabunin [Fri, 5 Dec 2014 14:48:28 +0000 (17:48 +0300)]
Build separate world-like iOS framework for contrib
Alexander Alekhin [Fri, 5 Dec 2014 12:55:54 +0000 (12:55 +0000)]
Merge pull request #3483 from wangyan42164:ocl_orb_improve
Alexander Alekhin [Fri, 5 Dec 2014 12:55:29 +0000 (12:55 +0000)]
Merge pull request #3482 from wangyan42164:ocl_brute_match_improve
Maksim Shabunin [Fri, 5 Dec 2014 12:33:53 +0000 (15:33 +0300)]
Fixed whitespace warnings in new tutorials
Andrey Kamaev [Fri, 5 Dec 2014 11:53:19 +0000 (14:53 +0300)]
Make Ctrl+S open save dialog on Gtk backend
Andrey Kamaev [Fri, 5 Dec 2014 11:06:01 +0000 (14:06 +0300)]
It turns opengl gheck is not needed for win ctrl+s
Yan Wang [Thu, 4 Dec 2014 07:48:23 +0000 (15:48 +0800)]
Use sincos instead of sin and cos.
Signed-off-by: Yan Wang <yan.wang@linux.intel.com>
Yan Wang [Thu, 4 Dec 2014 06:35:54 +0000 (14:35 +0800)]
Remove unnecesary operations and calculations in loop body.
Signed-off-by: Yan Wang <yan.wang@linux.intel.com>
Chuanbo Weng [Thu, 4 Dec 2014 03:15:13 +0000 (11:15 +0800)]
Remove unnecessary kercn limitation of 4.
When accessing global memory by DWORD4, memory bandwidth
can be fully utilized on Intel platform. This patch will
make more image format(e.g. 8UC4) be processed in DWORD4
by work-item. After applying this patch, 3 subcase of
./opencv_perf_core --gtest_filter=OCL_RepeatFixture_Repeat.Repeat/*
can be speedup on HD4000 graphics card with Beignet:
OCL_RepeatFixture_Repeat.Repeat/2, 64% improvement.
OCL_RepeatFixture_Repeat.Repeat/6, 50% improvement.
OCL_RepeatFixture_Repeat.Repeat/8, 56% improvement.
Signed-off-by: Chuanbo Weng <chuanbo.weng@intel.com>
TobyWanKenobi [Wed, 3 Dec 2014 16:13:11 +0000 (17:13 +0100)]
Re-edition of documentation
Re-edited the documentation again.
Maksim Shabunin [Wed, 3 Dec 2014 15:12:01 +0000 (18:12 +0300)]
Python test: added options to provide local data path
TobyWanKenobi [Wed, 3 Dec 2014 14:54:25 +0000 (15:54 +0100)]
Re-editing Documentation
Attempting to re-editing the documentation, to satisfy buildbot error
status.
TobyWanKenobi [Wed, 3 Dec 2014 14:40:12 +0000 (15:40 +0100)]
Changing tabs to spaces.
Edited previous code to change tabs into spaces.
Alexander Karsakov [Wed, 3 Dec 2014 13:01:03 +0000 (16:01 +0300)]
Fixed typo in description of Rect operations
fhsm [Tue, 2 Dec 2014 17:46:05 +0000 (12:46 -0500)]
White space and spelling in meanshift tutorial
Removed a leading whitespace from a parenthetical comment and corrected the spelling of "maybe"
TobyWanKenobi [Tue, 2 Dec 2014 14:15:06 +0000 (15:15 +0100)]
Add method in StatModel class, to load from String
Added a method "loadFromString" which is based on the "load" one. It
allow to directly pass the XML string which can be usefull and faster
when you have a huge file in a variable.
Maksim Shabunin [Tue, 2 Dec 2014 12:15:10 +0000 (15:15 +0300)]
Documentation: removed mention of Mat(IplImage*) constructor