Deepfake Detection on FaceForensics++ c23 (test)
99.8AUCFTCN
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| FTCN2022.01 | 99.8 | — | — | — | — | — | — | |
| RealForensics2022.01 | 99.8 | 99.1 | — | — | — | — | — | |
| TALLProcessing level=Video-level2026.05 | 99.8 | — | — | — | — | — | — | |
| LipForensics2022.01 | 99.7 | 98.8 | — | — | — | — | — | |
| SBIProcessing level=Frame-level2026.05 | 99.7 | — | — | — | — | — | — | |
| NACOProcessing level=Video-level2026.05 | 99.7 | — | — | — | — | — | — | |
| FTCNProcessing level=Video-level2026.05 | 99.7 | — | — | — | — | — | — | |
| SpInShieldProcessing level=Video-level2026.05 | 99.6 | — | — | — | — | — | — | |
| UCFProcessing level=Frame-level2026.05 | 99.5 | — | — | — | — | — | — | |
| AltFreezingProcessing level=Video-level2026.05 | 99.5 | — | — | — | — | — | — | |
| FakeSTormerProcessing level=Video-level2026.05 | 99.5 | — | — | — | — | — | — | |
| Xception2022.01 | 99.3 | 97 | — | — | — | — | — | |
| CNN-GRU2022.01 | 99.3 | 97 | — | — | — | — | — | |
| IIDProcessing level=Frame-level2026.05 | 99.3 | — | — | — | — | — | — | |
| CDFAProcessing level=Frame-level2026.05 | 99.3 | — | — | — | — | — | — | |
| DFD-FCGProcessing level=Video-level2026.05 | 99.3 | — | — | — | — | — | — | |
| ProDetProcessing level=Frame-level2026.05 | 99.2 | — | — | — | — | — | — | |
| CNN-aug2022.01 | 99.1 | 96.9 | — | — | — | — | — | |
| LSDAProcessing level=Frame-level2026.05 | 99 | — | — | — | — | — | — | |
| DCA EMNBackbone=DINOv3 ViT-L/16, Feature dimensionality=3,072 dims, Region=EMN (eyes-mouth-nose), Evaluation protocol=zero-shot2026.05 | 98.9 | — | — | — | — | — | — | |
| SLADDProcessing level=Frame-level2026.05 | 98.4 | — | — | — | — | — | — | |
| SLFProcessing level=Video-level2026.05 | 98.4 | — | — | — | — | — | — | |
| PTFDProcessing level=Video-level2026.05 | 98.4 | — | — | — | — | — | — | |
| DINOv2 ViT-L/14 DCA EMNBackbone=DINOv2 ViT-L/14, Region=EMN, Evaluation protocol=zero-shot2026.05 | 98.3 | — | — | — | — | — | — | |
| AFWProcessing level=Video-level2026.05 | 98 | — | — | — | — | — | — | |
| Face X-ray2022.01 | 97.8 | 78.4 | — | — | — | — | — | |
| Patch-based2022.01 | 97.2 | 92.6 | — | — | — | — | — | |
| A1 multi-region: mouth + skinBackbone=DINOv3 ViT-L/16, Region=mouth + skin, Evaluation protocol=zero-shot2026.05 | 96.9 | — | — | — | — | — | — | |
| A1 single region: mouthBackbone=DINOv3 ViT-L/16, Region=mouth, Evaluation protocol=zero-shot2026.05 | 96.3 | — | — | — | — | — | — | |
| A1 single region: noseBackbone=DINOv3 ViT-L/16, Region=nose, Evaluation protocol=zero-shot2026.05 | 95.8 | — | — | — | — | — | — | |
| SICATrain Set=FaceForensics++ (c23), Implementation=This paper2026.02 | 95.5 | — | 83.44 | 99.46 | 98.66 | 98.55 | 97.39 | |
| UCFTrain Set=FaceForensics++ (c23), Implementation=Original2026.02 | 94.92 | — | 83.99 | 98.83 | 98.4 | 98.96 | 94.41 | |
| A1 single region: eyesBackbone=DINOv3 ViT-L/16, Region=eyes, Evaluation protocol=zero-shot2026.05 | 94.7 | — | — | — | — | — | — | |
| EffortTrain Set=FaceForensics++ (c23), Implementation=Re-implementation2026.02 | 94.52 | — | 80.62 | 99.28 | 98.39 | 98.06 | 96.25 | |
| XceptionTrain Set=FaceForensics++ (c23), Implementation=Original2026.02 | 94.13 | — | 82.61 | 97.99 | 97.85 | 98.33 | 93.85 | |
| F3NetTrain Set=FaceForensics++ (c23), Implementation=Original2026.02 | 94.12 | — | 82.71 | 97.93 | 97.96 | 98.44 | 93.54 | |
| FFDTrain Set=FaceForensics++ (c23), Implementation=Original2026.02 | 93.97 | — | 82.37 | 98.03 | 97.84 | 98.53 | 93.06 | |
| CORETrain Set=FaceForensics++ (c23), Implementation=Original2026.02 | 93.9 | — | 81.94 | 97.87 | 98.03 | 98.23 | 93.39 | |
| RecceTrain Set=FaceForensics++ (c23), Implementation=Original2026.02 | 93.82 | — | 81.9 | 97.97 | 97.79 | 97.85 | 93.57 | |
| SPSLTrain Set=FaceForensics++ (c23), Implementation=Original2026.02 | 93.67 | — | 81.74 | 97.81 | 97.54 | 98.29 | 92.99 | |
| EfficientB4Train Set=FaceForensics++ (c23), Implementation=Original2026.02 | 93.54 | — | 81.5 | 97.57 | 97.58 | 97.97 | 93.08 | |
| X-rayTrain Set=FaceForensics++ (c23), Implementation=Original2026.02 | 93.5 | — | 79.25 | 97.94 | 98.72 | 98.71 | 92.9 | |
| SRMTrain Set=FaceForensics++ (c23), Implementation=Original2026.02 | 93.16 | — | 81.14 | 97.33 | 96.96 | 97.4 | 92.95 | |
| FaRL ViT-B/16 backbone + EMN pipelineBackbone=FaRL ViT-B/16, Region=EMN, Evaluation protocol=zero-shot2026.05 | 93 | — | — | — | — | — | — | |
| SiaTrain Set=FaceForensics++ (c23), Implementation=Re-implementation2026.02 | 90.32 | — | 73.67 | 96.21 | 94.92 | 95.37 | 91.43 | |
| DINOv3 CLS globalBackbone=DINOv3 ViT-L/16, Region=global CLS, Evaluation protocol=zero-shot2026.05 | 88.4 | — | — | — | — | — | — | |
| FWATrain Set=FaceForensics++ (c23), Implementation=Original2026.02 | 85.06 | — | 73.57 | 92.1 | 90 | 88.43 | 81.2 | |
| CNN-AugTrain Set=FaceForensics++ (c23), Implementation=Original2026.02 | 84.04 | — | 78.46 | 90.48 | 87.88 | 90.26 | 73.13 | |
| CapsuleTrain Set=FaceForensics++ (c23), Implementation=Original2026.02 | 81.76 | — | 70.4 | 86.69 | 86.34 | 87.34 | 78.04 | |
| MesoIncepTrain Set=FaceForensics++ (c23), Implementation=Original2026.02 | 75.69 | — | 72.78 | 85.42 | 80.87 | 74.21 | 65.17 | |
| Meso4Train Set=FaceForensics++ (c23), Implementation=Original2026.02 | 61.02 | — | 59.2 | 67.71 | 61.7 | 59.46 | 57.01 | |
| Cross-region Gram FrobeniusBackbone=DINOv3 ViT-L/16, Evaluation protocol=zero-shot2026.05 | 56 | — | — | — | — | — | — | |
| CADDMVenue=CVPR’23, Re-implemented=true2026.03 | — | 97.16 | — | — | — | — | — | |
| COREVenue=CVPRW’22, Re-implemented=true2026.03 | — | 96.61 | — | — | — | — | — | |
| EfficientNetVenue=ICML’19, Re-implemented=true2026.03 | — | 96.67 | — | — | — | — | — | |
| EffortVenue=ICML’25, Re-implemented=true2026.03 | — | 98.85 | — | — | — | — | — | |
| Face X-rayVenue=CVPR’20, Re-implemented=true2026.03 | — | 95.72 | — | — | — | — | — | |
| FIA-USAVenue=NeurIPS’25, Re-implemented=true2026.03 | — | 98.71 | — | — | — | — | — | |
| FoCusVenue=TIFS’242026.03 | — | 99.15 | — | — | — | — | — | |
| FreqBlender*Venue=NeurIPS’24, Re-implemented=true, Focuses on augmenting real images=true2026.03 | — | 93.29 | — | — | — | — | — | |
| GRUVenue=CVPR’242026.03 | — | 98.4 | — | — | — | — | — | |
| ProDetVenue=NeurIPS’242026.03 | — | 95.91 | — | — | — | — | — | |
| Qiao et al.Venue=TPAMI’242026.03 | — | 99 | — | — | — | — | — | |
| RDBC2026.03 | — | 99.41 | — | — | — | — | — | |
| RECCEVenue=CVPR’22, Re-implemented=true2026.03 | — | 96.95 | — | — | — | — | — | |
| SBI*Venue=CVPR’22, Re-implemented=true, Focuses on augmenting real images=true2026.03 | — | 85.16 | — | — | — | — | — | |
| Two-branch2022.01 | — | 99.1 | — | — | — | — | — | |
| UCFVenue=ICCV’23, Re-implemented=true2026.03 | — | 97.16 | — | — | — | — | — | |
| VLFFDVenue=CVPR’25, Re-implemented=true2026.03 | — | 98.64 | — | — | — | — | — |