Referee: Reference-aware Audiovisual Deepfake Detection

ArXi:2510.27475v2 Announce Type: replace Deepfakes generated by advanced generative models have rapidly posed serious threats, yet existing audiovisual deepfake detection approaches struggle to generalize to unseen manipulation methods. To address this, we propose a novel reference-aware audiovisual deepfake detection method, called Referee to capture fine-grained identity discrepancies.