Facial Forgery Detection on FaceForensics++ Hidden Benchmark low quality (test)
96.36Precision (DF)XceptionNet
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| XceptionNetInput Region=Face Region2019.01 | 96.36 | 86.86 | 90.29 | 80.67 | 52.4 | 70.1 | |
| MesoNetInput Region=Face Region2019.01 | 87.27 | 56.2 | 61.17 | 40.67 | 72.6 | 66 | |
| Cozzolino et al.Input Region=Face Region2019.01 | 85.45 | 67.88 | 73.79 | 78 | 34.4 | 55.2 | |
| Rahmouni et al.Input Region=Face Region2019.01 | 85.45 | 64.23 | 56.31 | 60.07 | 50 | 58.1 | |
| Bayar and StammInput Region=Face Region2019.01 | 84.55 | 73.72 | 82.52 | 70.67 | 46.2 | 61.6 | |
| Xcept. Full ImageInput Region=Full Image2019.01 | 74.55 | 75.91 | 70.87 | 73.33 | 51 | 62.4 | |
| Steg. FeaturesInput Region=Face Region2019.01 | 73.64 | 73.72 | 68.93 | 63.33 | 34 | 51.8 |