Audio-Visual Deepfake Detection on AV-Deepfake1M Trimmed
0.835AUCAVH-Align
Evaluation results
| Method | Links | ||
|---|---|---|---|
| AVH-AlignModality=AV, Train type=unsup., Train data=VoxCeleb22024.11 | 0.835 | 0.935 | |
| AVH-Align/supModality=AV, Train type=sup., Train data=AV1M2024.11 | 0.831 | 0.944 | |
| SpeechForensicsModality=AV, Train type=unsup., Train data=VoxCeleb22024.11 | 0.682 | 0.835 | |
| AVH-Align/supModality=AV, Train type=sup., Train data=FAVC2024.11 | 0.636 | 0.822 | |
| AVADModality=AV, Train type=unsup., Train data=LRS2024.11 | 0.543 | 0.763 |