Hi @taketwo,

have you considered adding CUDA support to radical? What I mean with that is to give the possibility to pass device memory pointers (maybe cv::GpuMat) to the radiometric and vignetting response interface functions.

When you handle your images in device memory it is really unfortunate (mainly in terms of speed) that you have to copy back and forth between CPU and GPU when you want to use radical functionality.

I think it would be an awesome feature. :)

Best regards,
Tim