Research team from Tokyo University and the MPI-INF has introduced a new approach for detecting manipulated face videos. Instead of searching for pixel-level inconsistencies, their model predicts how a person’s face should move based on their speech. #CVPR2026 ↪️ mpi-inf.link/EACVPR2026
Researchers develop new method for detecting manipulated videos
The AI-based approach turns video footage into a detector for Deepfakes of a specific person.
mpi-inf.link