Deepfake Detection on DFDC (Demographic Analysis)
1.76Gender FPROurs
Evaluation results
| Method | Links | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| OursBackbone=Xception2025.11 | 1.76 | 3.67 | 57.94 | 17.93 | 7.58 | 52.33 | 37.37 | 11.92 | 39.03 | 61.86 | |
| RSEF-FDDBackbone=Xception2025.11 | 2.09 | 3.83 | 57.9 | 12.24 | 7.74 | 50.27 | 38.89 | 12.18 | 37.72 | 59.08 | |
| PG-FDDBackbone=Xception2025.11 | 2.35 | 2.57 | 57.71 | 10.1 | 11.49 | 50.53 | 32.16 | 17.71 | 37.42 | 59.64 | |
| FairadapterBackbone=ViT-L/142025.11 | 2.39 | 4.57 | 55.88 | 32.49 | 26.77 | 52.27 | 74.1 | 31.53 | 31.39 | 59.25 | |
| DAW-FDDBackbone=Xception2025.11 | 3.6 | 3.67 | 55.88 | 22.36 | 9.72 | 47.77 | 46.8 | 12.69 | 36.63 | 56.9 | |
| DAG-FDDBackbone=Xception2025.11 | 5.29 | 6.68 | 57.41 | 9.26 | 12.51 | 45.91 | 45.3 | 12.92 | 36.42 | 59.13 | |
| OriBackbone=Xception2025.11 | 8.67 | 6.7 | 54.56 | 19.41 | 7.99 | 47.02 | 53.42 | 17.34 | 36.96 | 56.13 |