Code for the paper "DeepRemaster: Temporal Source-Reference Attention Networks for Comprehensive Video Enhancement". http://iizuka.cs.tsukuba.ac.jp/projects/remastering/
I love this paper, and the work you've put into this. Are you interested in training the model with video degradation, so that it could say restore / improve old video, or even poor mpeg encoding and blocking? One could flag --type film or video for example.
I love this paper, and the work you've put into this. Are you interested in training the model with video degradation, so that it could say restore / improve old video, or even poor mpeg encoding and blocking? One could flag --type film or video for example.