Frame-level Deepfake Detection on DiffSwap
86.98AUC (%)DiffusionFake (ViT-B)
Evaluation results
| Method | Links | ||
|---|---|---|---|
| DiffusionFake (ViT-B)Backbone=ViT-B, Training Dataset=FF++(HQ)2024.10 | 86.98 | 21.32 | |
| DiffusionFake (En-b4)Backbone=EfficientNet-B4, Training Dataset=FF++(HQ)2024.10 | 82.02 | 25.55 | |
| DCLTraining Dataset=FF++(HQ), Reproduced Status=false2024.10 | 80.21 | 27.37 | |
| MATTraining Dataset=FF++(HQ), Reproduced Status=true2024.10 | 79.93 | 27.77 | |
| ViT-BBackbone=ViT-B, Training Dataset=FF++(HQ), Reproduced Status=true2024.10 | 78.5 | 28.14 | |
| GFFTraining Dataset=FF++(HQ), Reproduced Status=true2024.10 | 78.38 | 28.15 | |
| LTWTraining Dataset=FF++(HQ), Reproduced Status=false2024.10 | 77.95 | 29.01 | |
| RECCETraining Dataset=FF++(HQ), Reproduced Status=true2024.10 | 77.59 | 29.38 | |
| EfficientNet-B4 (En-b4)Backbone=EfficientNet-B4, Training Dataset=FF++(HQ), Reproduced Status=true2024.10 | 77.38 | 29.44 | |
| F3-NetTraining Dataset=FF++(HQ), Reproduced Status=true2024.10 | 76.89 | 30.83 | |
| CADDMTraining Dataset=FF++(HQ), Reproduced Status=true2024.10 | 75.58 | 31.01 | |
| SBITraining Dataset=FF++(HQ), Reproduced Status=true2024.10 | 75.2 | 31.49 | |
| XceptionBackbone=Xception, Training Dataset=FF++(HQ), Reproduced Status=false2024.10 | 74.25 | 32.04 |