Which deep learning technique would you use for video compression?
Sigiloso
I would choose architecture similar to that of VAE to find the embeddings into lower dimensional space. This would reduce the dimensions of the frames and eventually compress the video.