A single video super-resolution GAN for multiple downsampling operators based on pseudo-inverse image formation models | Publicación