Compare performances of algorithms on Deepfake Detection.
Conf. | Methods | FaceForensics++ (raw) | FaceForensics++ (c23) | FaceForensics++ (c40) | UADFV | DeepfakeTIMIT-LQ | DeepfakeTIMIT-HQ | FaceForensic | DFFD | celeb-DF | DFDC |
Accuracy | Accuracy/AUC | Accuracy/AUC | AUC | AUC | AUC | Accuracy | AUC/EER | AUC/EER | AUC/EER/logloss | ||
CVPR21 | Multi-attentional Deepfake Detection | - | 97.60/99.29 | 88.69/90.40 | - | - | - | - | - | - | -/-/0.1679 |
ECCV20 | Two-branch Recurrent Network | - | 96.43/99.12 | 86.34/91.10 | - | - | - | - | - | - | - |
FG20 | 3D ResNet | - | - | 83.86 | - | - | - | - | - | - | - |
3D ResNeXt | - | - | 85.14 | - | - | - | - | - | - | - | |
I3D | - | - | 87.43 | - | - | - | - | - | - | - | |
CVPR20 | VGG16 | - | - | - | - | - | - | - | 96.95/8.43 | - | - |
VGG16+MAM | - | - | - | - | - | - | - | 99.67/2.66 | - | - | |
ICCV19 | XceptionNet | 99.26 | 95.73/92.50 | 81.00/86.75 | 96.8 | - | - | - | 99.61/2.88 | 83.60/28.55 | 91.17/17.55/- |
ICASSP19 | HeadPose | - | - | - | 89 | - | - | - | - | - | - |
CVPRW19 | ResNet50-finetune | - | - | - | 97.4 | 99.9 | 93.2 | - | - | - | - |
BTAS19 | Y-shaped Autorncoder | - | - | - | - | - | - | 92.77 | - | - | - |
CVPRW19 | FWA | - | - | - | 97.4 | - | - | - | - | - | - |
ICASSP19 | capsule | 99.52 | - | - | 99.9 | - | - | - | - | 82.46/24.29 | 87.45/21.39/- |
WIFS18 | MesoNet | 95.23 | 83.1 | 70.47 | 84.3 | 87.8 | 68.4 | 93.4 | - | - | - |
arxiv18 | ForensicTransfer | - | - | - | - | - | - | 92.6 | - | - | - |
CVPRW17 | two-stream NN | - | - | - | 85.1 | 83.5 | 73.5 | - | - | - | - |