Merge remote-tracking branch 'origin/2.4' into merge-2.4

Conflicts:
modules/core/src/gl_core_3_1.cpp
modules/core/src/opencl/convert.cl
modules/cudaimgproc/src/cuda/canny.cu
modules/cudastereo/perf/perf_stereo.cpp
modules/gpu/perf/perf_imgproc.cpp
modules/gpu/test/test_denoising.cpp
modules/ocl/src/opencl/imgproc_resize.cl
samples/cpp/Qt_sample/main.cpp

Merge pull request #2337 from pmoulon:master

Merge pull request #2328 from rohitgirdhar:bugfix_blobdetector_kpt_radius

Merge pull request #2330 from rohitgirdhar:linux_tutorial_make_j8

Merge pull request #2342 from yashdv:android_doc_fix

Merge pull request #2348 from ilya-lavrenov:typo

Merge pull request #2351 from yashdv:BOW_assert_error

Merge pull request #2349 from jet47:fix-bug-3552

Merge pull request #2344 from kirill-kornyakov:replace-gittip-button

Fixed CV_ASSERT condition in the code for bag of words

fix bug #3552:

replace std::swap with own code

typo

Replaced Gittip button

fixed doc/tutorials/introduction/android_binary_package/dev_with_OCV_on_Android.rst

Merge pull request #2336 from SpecLad:useOclChanged

Merge pull request #2335 from jet47:gpu-alphacomp-test

Merge pull request #2326 from vpisarev:ocl_ch3

Merge pull request #2323 from stuart8c:issue2601_2.4_fix_16bit_tiff_reading_on_big_endian_host

Fix a typo error

Fix a typo error

Merge pull request #2329 from ilya-lavrenov:macosx_ogl

Deleted an assignment to a variable that doesn't exist.

temporary disable performance test for alphaComp function

Merge pull request #2315 from ilya-lavrenov:nlmeans_perf

Merge pull request #1735 from KonstantinMatskevich:umat_tests

Merge pull request #2062 from cook:bug3440

Fix build of libtiff on big endian host due to defined but empty WORDS_BIGENDIAN macro

some more fixes

Use CMAKE_CFG_INTDIR meta-variable to get the library output directory of the opencv_matlab.

Merge pull request #2333 from asmorkalov:java_dead_code_fix

Merge pull request #2332 from jet47:gpu-fixes

Dead code removed as this cannot be null in Java.

Merge pull request #2331 from asmorkalov:android_native_activity_finish

fixed OpenCL kernel flag in resize (thanks to Ilya L)

temporary disable perf test for StereoBeliefPropagation

increase epsilon for AlphaComp sanity test for integer input

Application pause/resume fix for Android sample NativeActivity.

Fix cmake detection of build platform endianness

Improve comments to indicate actual usage of WORDS_BIGENDIAN where
it is tested with #ifdef rather than #if

update

Merge pull request #2304 from mattvenn:master

-j8 in linux_installation tutorial

Merge pull request #2318 from gongzg:2.4

Merge pull request #2319 from m3d:patch-1

Merge pull request #2324 from vpisarev:fix_facedetect

added imgpoints back

Some tests for UMat

enabled OpenGL on MacOSX

bugfix: blobdetector was putting incorrect diameter value in keypoints

tab => spaces

3-channel support in OpenCL kernels for setTo, resize, warpAffine and warpPerspective

fixed bug reported by Irina K. Also tried to fix LBP detector regression in quality (probably it did not help but still ...)

Fix Windows build of grfmt_tiff.cpp by using libtiff's integer types.

Merge pull request #2322 from berak:b_cc_read

Merge pull request #2320 from asmorkalov:ocv_readme_license_install

Merge pull request #2321 from asmorkalov:android_run_all_tests

Merge pull request #2317 from blueskymonster:master

undo changes of cv2.cv.* functions/constants

opencv_run_all_tests.sh implemented for Android SDK.

Fix reading of 16-bit TIFF images on big endian host.

Use correct integer types for arguments to TIFFGetField to avoid corruption
of values and failed loads of TIFF file when using cv::imread().

Added test where both big and little endian TIFF files are read using imread().

Fixed build of 3rdparty libtiff on big endian hosts.

Reduced memory required during decode_tile16384x16384 test by not converting
large grayscale test image to color image during read.

LICENSE and README files installation rules added.

added 3-channel support to arithmetic operations

removed return images and fixed cv2 names

Updated DisplayImage.cpp program to compile

CV_WINDOW_AUTOSIZE -> WINDOW_AUTOSIZE

Workaround a LLVM/Clang 3.3 bug.

The LLVM/Clang 3.3 has a bug when compile a cl kernel with
assignment of a scalar to a vector data type. This patch
could work around this bug.

Signed-off-by: Zhigang Gong <zhigang.gong@intel.com>

added performance tests for NLMeans

Merge pull request #2311 from SvenWe:master

Merge pull request #2313 from m3d:patch-1

Merge pull request #2314 from ilya-lavrenov:typo

Merge pull request #1851 from dneelyep:patch-1

typo

Merge pull request #2310 from SpecLad:merge-2.4

Merge remote-tracking branch 'origin/2.4' into merge-2.4

Conflicts:
CMakeLists.txt
cmake/OpenCVGenAndroidMK.cmake
cmake/templates/OpenCV.mk.in
doc/tutorials/viz/creating_widgets/creating_widgets.rst
doc/tutorials/viz/launching_viz/launching_viz.rst
doc/tutorials/viz/table_of_content_viz/images/image_effects.png
doc/tutorials/viz/transformations/transformations.rst
doc/tutorials/viz/widget_pose/widget_pose.rst
modules/core/include/opencv2/core/affine.hpp
modules/core/include/opencv2/core/core.hpp
modules/gpu/perf/perf_imgproc.cpp
modules/gpu/src/cuda/canny.cu
modules/gpu/src/cuda/generalized_hough.cu
modules/gpu/src/generalized_hough.cpp
modules/gpu/src/imgproc.cpp
modules/gpu/test/test_color.cpp
modules/gpu/test/test_core.cpp
modules/gpu/test/test_gpumat.cpp
modules/gpu/test/test_hough.cpp
modules/nonfree/CMakeLists.txt
modules/nonfree/include/opencv2/nonfree/gpu.hpp
modules/nonfree/perf/perf_gpu.cpp
modules/nonfree/src/cuda/surf.cu
modules/nonfree/src/precomp.hpp
modules/nonfree/src/surf_gpu.cpp
modules/nonfree/test/test_gpu.cpp
modules/ocl/perf/perf_haar.cpp
modules/stitching/CMakeLists.txt
modules/stitching/include/opencv2/stitching/detail/matchers.hpp
modules/stitching/include/opencv2/stitching/detail/seam_finders.hpp
modules/stitching/include/opencv2/stitching/detail/warpers.hpp
modules/stitching/include/opencv2/stitching/warpers.hpp
modules/stitching/src/blenders.cpp
modules/stitching/src/matchers.cpp
modules/stitching/src/precomp.hpp
modules/stitching/src/seam_finders.cpp
modules/stitching/src/stitcher.cpp
modules/stitching/src/warpers.cpp
modules/viz/doc/widget.rst
modules/viz/include/opencv2/viz/types.hpp
modules/viz/include/opencv2/viz/viz3d.hpp
modules/viz/include/opencv2/viz/widget_accessor.hpp
modules/viz/src/precomp.hpp
modules/viz/src/shapes.cpp
modules/viz/src/vizcore.cpp
modules/viz/src/vtk/vtkCloudMatSink.h
modules/viz/src/vtk/vtkCloudMatSource.h
modules/viz/test/test_precomp.hpp
modules/viz/test/tests_simple.cpp
samples/android/tutorial-4-cuda/CMakeLists.txt
samples/android/tutorial-4-cuda/jni/Android.mk
samples/android/tutorial-4-cuda/src/org/opencv/samples/tutorial4/Tutorial4Activity.java
samples/cpp/stitching_detailed.cpp
samples/cpp/tutorial_code/viz/creating_widgets.cpp
samples/cpp/tutorial_code/viz/launching_viz.cpp
samples/cpp/tutorial_code/viz/transformations.cpp
samples/cpp/tutorial_code/viz/widget_pose.cpp

fixed spelling and removed return images

grabcut.py sample: Allow drawing rect from bottom/right

The grabcut sample script isn't able to create rectangles correctly that are
drawn from e.g. bottom-right to top-left.
This small commit uses four min() calls to determine the correct
top-left point of the rectangle.
It also removes some unnecessary white spaces before colons.

removed whitespace at end of line 96

Merge pull request #2308 from jet47:gpu-tests-timeout

Merge pull request #2309 from Nerei:fixed_cl_exe_hang_with_eigen

Fixed Cl.exe hang for VS2008 x64 with Eigen::exp() function

decrease input size for several gpu tests to fix "timed out" error:

* BruteForceNonLocalMeans
* OpticalFlowBM

proper initialization for ccimpl in bool CascadeClassifier::read(const FileNode

fixed whitespace

fixed examples for opencv 2.4.8

Merge pull request #2303 from jet47:gpu-canny-fix

Merge pull request #2295 from alalek:opencl_align_rw_buffers

Merge pull request #2294 from jet47:fix-gpu-issues

Merge pull request #2296 from apavlenko:fix_coverity_scan

fixed several bugs in CUDA Canny implementation:

* out of border access in edgesHysteresisLocalKernel
* incorrect usage of atomicAdd

Merge pull request #2301 from jet47:fix-nonfree-build

Merge pull request #2297 from asmorkalov:ocv_testdata_install_perms_fix

Revert "disable gpu Canny and HoughCircles perf tests:"

This reverts commit dbce90692acd84fbf46bde4da4b1726049f42857.

ocl: update gftt

ocl: workaround for failures with misaligned host buffers

Implicit testdata directory permissions setup added.

Merge pull request #2300 from asmorkalov:android_no_cuda_not_armeabiv7a

fix nonfree module compilation without CUDA

dynamicuda module disabled in OpenCv.mk for all arches except armeabi-v7a.

Merge pull request #2251 from ilya-lavrenov:tapi_sqrboxfilter

Merge pull request #2287 from ilya-lavrenov:mat_step

Merge pull request #2289 from ilya-lavrenov:umat_expr

fixing issues from coverity scan 1167937, 1167938, 1167939, 1127248 (scan.coverity.com/projects/169?tab=Overview)