Modeling Biomechanical Constraint Violations for Language-Agnostic Lip-Sync Deepfake Detection
It addresses the problem of cross-language generalization in lip-sync deepfake detection for security and media forensics.
The paper identifies that generative models violate biomechanical constraints of orofacial articulation, producing measurable temporal lip jitter, and proposes BioLip, a lightweight detector using 64 perioral landmarks, achieving language-agnostic lip-sync deepfake detection.
Current lip-sync deepfake detectors rely on pixel-level artifacts or audio-visual correspondence, failing to generalize across languages because these cues encode data-dependent patterns rather than universal physical laws. We identify a more fundamental principle: generative models do not enforce the biomechanical constraints of authentic orofacial articulation, producing measurably elevated temporal lip variance -- a signal we term temporal lip jitter -- that is empirically consistent across the speaker's language, ethnicity, and recording conditions. We instantiate this principle through BioLip, a lightweight framework operating on 64 perioral landmark coordinates extracted by MediaPipe.