DART: A Degradation-Aware Recurrent Transformer for Archival Film Restoration

DART is a degradation-aware recurrent transformer for restoring old film by explicitly processing damage rather than just reconstructing pixels. It predicts and propagates a soft defect mask over time, using this mask to guide temporal fusion and condition the restoration network on both where the damage is and how severe it is, resulting in cleaner, temporally consistent restorations.

Open the full episode

More from AI Radio