Face Forgery Detection on Celeb-DF v1 (frame-level AUC)
91.6Frame-level AUCSFAM
Evaluation results
| Method | Links | |
|---|---|---|
| SFAMTraining Dataset=FaceForensics++, Evaluation Level=frame-level2026.04 | 91.6 | |
| Forensics AdapterTraining Dataset=FaceForensics++, Evaluation Level=frame-level2026.04 | 91.4 | |
| EffortTraining Dataset=FaceForensics++, Evaluation Level=frame-level2026.04 | 89.2 | |
| UCFTraining Dataset=FaceForensics++, Evaluation Level=frame-level2026.04 | 81.1 | |
| Efficientnet-B4Training Dataset=FaceForensics++, Evaluation Level=frame-level2026.04 | 76.7 | |
| XceptionTraining Dataset=FaceForensics++, Evaluation Level=frame-level2026.04 | 75.4 | |
| F3NetTraining Dataset=FaceForensics++, Evaluation Level=frame-level2026.04 | 75 | |
| RECCETraining Dataset=FaceForensics++, Evaluation Level=frame-level2026.04 | 73.4 | |
| CLIPTraining Dataset=FaceForensics++, Evaluation Level=frame-level2026.04 | 71.6 | |
| FFDTraining Dataset=FaceForensics++, Evaluation Level=frame-level2026.04 | 70.9 |