finished with NV12 support for D3D11-interop. Now, if texture is in NV12 format then it will be converted to/from BGR UMat.

fixed issues with ocl nv12 cvt kernel

finisged ocl nv12-to-rgba kernel, update dx-interop samples. (ocl rgba-to-nv12 kernel will be added later)

an attempt to fix build issue

fix for non opencl build issue

fix typo

fix compilation warnings

fix compile issue for Mac (OpenCL)

add convertion from rgba to nv12 (still need to debug kernel)

remove empty line at the EOF

fixed compilation warning

misplaced CV_WRAP statement lead to not exposing EM::getCovs() to python/java/matlab wrappers

add a CV_OUT to return the covs properly

Now StereoSGBM behaves similarly to StereoBM in terms of ROI handling.

New mode is approximately 2-3 times faster than MODE_SGBM
with minimal degradation in quality and uses universal
HAL intrinsics. A performance test was added. The accuracy
test was updated to support the new mode.

Signed-off-by: NYan Wang <yan.wang@linux.intel.com>

It could reduce 7 multiplications every iteration.
Signed-off-by: NYan Wang <yan.wang@linux.intel.com>

rewrite & change convertFromGLBuffer() & convertToGLBuffer() into acquireGLBuffer() & releaseGLBuffer(), respectively

opengl sample: added buffer support

tested and fixed buffer support on Windows

change glFlush() call to glFinish()

added UMat::release() call; fixed functions' names

adopted & implemented API suggestion(s) from Alexander

fixed unreachable code warning

added more info to the mapGLBuffer/unmapGLBuffer description

pos should never exceed strlen, else the overflow check 'if (!len)' will wrongfully pass

1. Remove uncessary index calculation.
2. Use mad/mad24 as possible.
Signed-off-by: NYan Wang <yan.wang@linux.intel.com>

by this new function we can set libjpeg "scale_denom" parameter and load jpeg images scaled 1/2 - 1/4 - 1/8
other image formats resized after loading

compute fps changed (now only interop counts). fixed d3d10 and d3d11 print info method (there were issues with texture mapping)

Fixed polarity of ImageRegistration mode in setDepthGeneratorProperty function of CvCapture_OpenNI2 class

it provides more expectable results to end-users.
based on feedback by Alexander Alekhin

a separate function instead of an overload not to change the ABI.

rename VideoCapture paramter 'device' to 'index' in CPP to reflect that
it allows specifying the API.
update comments to explain how to specify the API.

http://code.opencv.org/issues/4462