Deepfake Detection on FaceForensics++ DF c23
100AUROCPCL + I2G
Evaluation Results
| Method | Links | |
|---|---|---|
| PCL + I2GType of Model=Closed-set, Training Data=FF++c232025.03 | 100 | |
| S-MIL-TType of Model=Closed-set, Training Data=FF++c232025.03 | 99.84 | |
| MILType of Model=Closed-set, Training Data=FF++c232025.03 | 99.51 | |
| XN-avgType of Model=Closed-set, Training Data=FF++c232025.03 | 99.38 | |
| ProposedType of Model=Open-set, Training Data=FF++c23, FS excluded2025.03 | 99.15 | |
| Face X-rayType of Model=Closed-set, Training Data=FF++c232025.03 | 99.12 | |
| ProposedType of Model=Open-set, Training Data=FF++c23, NT excluded2025.03 | 98.96 | |
| ProposedType of Model=Open-set, Training Data=FF++c232025.03 | 98.96 | |
| ProposedType of Model=Open-set, Training Data=FF++c23, F2F excluded2025.03 | 98.63 |