I have an OpenCL implementation of imfilter that I've been wanting to port over here for a while. Unfortunately, I can't understand the complex dispatch in imfilter.jl to contribute it myself. I appreciate if you can give some instructions. Also feel free to port it if you feel it is easy: https://github.com/JuliaEarth/ImageQuilting.jl/blob/master/src/imfilter_gpu.jl
Even though OpenCL CLFFT only works with specific sizes of images that are multiple of prime numbers, the code I've implemented above does the necessary padding and works with any image size.
I have an OpenCL implementation of
imfilter
that I've been wanting to port over here for a while. Unfortunately, I can't understand the complex dispatch inimfilter.jl
to contribute it myself. I appreciate if you can give some instructions. Also feel free to port it if you feel it is easy: https://github.com/JuliaEarth/ImageQuilting.jl/blob/master/src/imfilter_gpu.jlEven though OpenCL CLFFT only works with specific sizes of images that are multiple of prime numbers, the code I've implemented above does the necessary padding and works with any image size.