A research implementation extending RA-Det from image-level representation analysis to temporal video representation analysis for modern deepfake detection.
machine-learning pytorch representation-learning image-forensics temporal-modeling vision-transformer video-deepfake synthetic-video video-forensics perturbation-learning representation-discrepancy robust-feature-learning ra-det
-
Updated
Aug 5, 2026 - Python