more optimization for warpAffine and warpPerspective
authorLi Peng <peng.li@intel.com>
Tue, 29 Nov 2016 07:28:53 +0000 (15:28 +0800)
committerLi Peng <peng.li@intel.com>
Wed, 30 Nov 2016 07:43:41 +0000 (15:43 +0800)
commit2ca5a7e862b9f7c4de90ddef73ae7c12c8981866
tree90942e98fa972d3f90c68c7d4ac5700c6f15a955
parentc47267ef7f118ac6f78cce45727a2c57006f3f95
more optimization for warpAffine and warpPerspective

Add new OpenCL kernels for bicubic interploation, it is 20% faster
than current warp image kernel with bicubic interploation.

Signed-off-by: Li Peng <peng.li@intel.com>
modules/imgproc/src/imgwarp.cpp
modules/imgproc/src/opencl/warp_transform.cl
modules/imgproc/test/ocl/test_warp.cpp