I've managed to implement "overlay" filter to overlay the png to the video, but I need to use hardware to speed up the processing of the overlay. So I want to use "overlay_cuda" filter in C++. But I couldn't find the example of using the "overlay_cuda" filter. And the use of "overlay_cuda" filter seems be different from the use of "overlay" filter, So I want to the how to use overlay_cuda filter(FFmpeg API) in C / C++ ? Are there some examples? Thank you for help~