Deepfake Attribution on OW-DFA-40 Protocol-1
88.3All AccuracyCAL
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| CALKnowledge of number of novel forgery types (KU)=known2025.12 | 88.3 | 90.3 | 91.5 | 76.5 | 82.2 | 72.4 | 98 | |
| CAL (w/o KU)Knowledge of number of novel forgery types (KU)=unknown2025.12 | 87.5 | 89.7 | 90.2 | 74.5 | 82.3 | 70.9 | 98.2 | |
| CDALKnowledge of number of novel forgery types (KU)=known2025.12 | 84.3 | 87.4 | 86 | 70 | 77.7 | 62.5 | 95.3 | |
| CPLKnowledge of number of novel forgery types (KU)=known2025.12 | 82.6 | 85.5 | 85.1 | 64.3 | 73.2 | 56.3 | 96.2 | |
| PALGCDKnowledge of number of novel forgery types (KU)=known2025.12 | 76.5 | 80 | 79 | 54.2 | 66.5 | 45.1 | 95.9 | |
| LPSKnowledge of number of novel forgery types (KU)=known2025.12 | 76.1 | 79 | 75.4 | 61.2 | 69.5 | 53.1 | 90.9 | |
| OwMatchKnowledge of number of novel forgery types (KU)=known2025.12 | 74.1 | 82.1 | 81.7 | 61.2 | 74.3 | 53.6 | 96.2 | |
| SimGCDKnowledge of number of novel forgery types (KU)=known2025.12 | 73.3 | 82.5 | 54.5 | 65 | 75 | 57.6 | 81.7 | |
| LegoGCDKnowledge of number of novel forgery types (KU)=known2025.12 | 70.9 | 76.3 | 57.6 | 61 | 69.2 | 52.9 | 82.6 | |
| ProtoGCDKnowledge of number of novel forgery types (KU)=known2025.12 | 62.9 | 79.7 | 37.7 | 59.8 | 71.1 | 51.3 | 66.8 |