-
Li Peng authored
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>
2ca5a7e8
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>