Deepfake Detection on DeepfakeTIMIT HQ
0.9998AUCAV-Lip-Sync+ with FE
Evaluation results
| Method | Links | ||
|---|---|---|---|
| AV-Lip-Sync+ with FEType=Multimodal, Modality=AV2023.11 | 0.9998 | — | |
| DSP-FWAType=Unimodal, Modality=V2023.11 | 0.997 | — | |
| POI-ForensicsType=Multimodal, Modality=AV2023.11 | 0.992 | — | |
| MCLType=Multimodal, Modality=AV2023.11 | 0.9909 | — | |
| AV-Lip-Sync+Type=Multimodal, Modality=AV2023.11 | 0.988 | — | |
| MDSType=Multimodal, Modality=AV2023.11 | 0.9687 | — | |
| AV-Lip-SyncType=Multimodal, Modality=AV2023.11 | 0.968 | — | |
| Emotions don't lieType=Multimodal, Modality=AV2023.11 | 0.949 | — | |
| Xception-c23Type=Unimodal, Modality=V2023.11 | 0.944 | — | |
| FWAType=Unimodal, Modality=V2023.11 | 0.932 | — | |
| VA-LogRegType=Unimodal, Modality=V2023.11 | 0.773 | — | |
| CapsuleType=Unimodal, Modality=V2023.11 | 0.744 | — | |
| Two-streamType=Unimodal, Modality=V2023.11 | 0.735 | — | |
| Xception-c40Type=Unimodal, Modality=V2023.11 | 0.705 | — | |
| Meso-4Type=Unimodal, Modality=V2023.11 | 0.684 | — | |
| VA-MLPType=Unimodal, Modality=V2023.11 | 0.621 | — | |
| Multi-taskType=Unimodal, Modality=V2023.11 | 0.553 | — | |
| Xception-rawType=Unimodal, Modality=V2023.11 | 0.54 | — | |
| HeadPoseType=Unimodal, Modality=V2023.11 | 0.532 | — | |
| CNN+LSTMResolution=128x1282026.05 | — | 87.6 | |
| Frequency DomainResolution=128x1282026.05 | — | 82.1 | |
| Optical FlowResolution=128x1282026.05 | — | 83.2 | |
| Proposed 3D CNNResolution=128x1282026.05 | — | 92.8 | |
| XceptionNetResolution=128x1282026.05 | — | 84.5 |