S3VD Self-Supervised Spatial Video Downsampling Loss: A Method for Training Video FPN Denoising Networks | Publicación