Deepfake Detection on DFD (test)
98.25AccuracyMARE
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| MARE2026.01 | 98.25 | 99.68 | — | — | — | |
| ExpDVenue=AAAI20242026.01 | 97.53 | 99.23 | — | — | — | |
| M2F2Venue=CVPR20252026.01 | 96.51 | 98.67 | — | — | — | |
| EfficientNetVenue=ICML20192026.01 | 95.63 | 98.14 | — | — | — | |
| SFICVenue=TIFS20242026.01 | 93.74 | 94.57 | — | — | — | |
| GTDetector=LSDA, Type=CVPR’242026.04 | 88 | — | — | 253.2 | — | |
| GTDetector=ViT-Base, Type=Backbones2026.04 | 86.7 | — | — | 330.2 | — | |
| Ground Truth (GT)Backbone=ViT-Base2026.04 | 86.7 | — | — | — | — | |
| GTDetector=Swin-Base, Type=Backbones2026.04 | 86.5 | — | — | 334.8 | — | |
| Ground Truth (GT)Backbone=Swin-Base2026.04 | 86.5 | — | — | — | — | |
| GTDetector=Wavelet, Type=WACV’252026.04 | 85.7 | — | — | 615 | — | |
| GTDetector=NPR, Type=CVPR’242026.04 | 85.5 | — | — | 177.9 | — | |
| GTDetector=Swin-Small, Type=Backbones2026.04 | 85 | — | — | 189.2 | — | |
| GTDetector=UCF, Type=ICCV’232026.04 | 84.9 | — | — | 91.6 | — | |
| GTDetector=ViT-Small, Type=Backbones2026.04 | 84.5 | — | — | 84.1 | — | |
| GTDetector=DeiT-Base, Type=Backbones2026.04 | 82.3 | — | — | 330.2 | — | |
| Ground Truth (GT)Backbone=DeiT-Base2026.04 | 82.3 | — | — | — | — | |
| GTDetector=DeiT-Small, Type=Backbones2026.04 | 79.4 | — | — | 84.1 | — | |
| DeFakeQDetector=Wavelet, Type=WACV’252026.04 | 78.9 | — | — | 84.8 | 86.2 | |
| DeFakeQDetector=LSDA, Type=CVPR’242026.04 | 78.4 | — | — | 26.6 | 89.5 | |
| DeFakeQDetector=ViT-Base, Type=Backbones2026.04 | 78.1 | — | — | 31.2 | 90.6 | |
| DeFakeQTechnology Category=Quantization, Backbone=ViT-Base2026.04 | 78.1 | — | — | — | — | |
| GTDetector=DeiT-Tiny, Type=Backbones2026.04 | 77.8 | — | — | 23.7 | — | |
| DeFakeQDetector=NPR, Type=CVPR’242026.04 | 77.6 | — | — | 19.7 | 88.9 | |
| DeFakeQDetector=ViT-Small, Type=Backbones2026.04 | 77.3 | — | — | 8.2 | 90.2 | |
| DeFakeQDetector=UCF, Type=ICCV’232026.04 | 77.3 | — | — | 9.1 | 90 | |
| DeFakeQDetector=Swin-Base, Type=Backbones2026.04 | 77.1 | — | — | 31.8 | 90.5 | |
| DeFakeQTechnology Category=Quantization, Backbone=Swin-Base2026.04 | 77.1 | — | — | — | — | |
| DeFakeQDetector=Swin-Small, Type=Backbones2026.04 | 76.9 | — | — | 18.1 | 90.4 | |
| DeFakeQDetector=DeiT-Base, Type=Backbones2026.04 | 74.5 | — | — | 31.7 | 90.4 | |
| DeFakeQTechnology Category=Quantization, Backbone=DeiT-Base2026.04 | 74.5 | — | — | — | — | |
| DeFakeQDetector=DeiT-Small, Type=Backbones2026.04 | 72.1 | — | — | 7.5 | 91.1 | |
| DeFakeQDetector=DeiT-Tiny, Type=Backbones2026.04 | 69.3 | — | — | 2.2 | 90.7 | |
| FIMA-QDetector=Swin-Small, Type=Backbones2026.04 | 63.4 | — | — | 17 | 91 | |
| FIMA-QDetector=ViT-Small, Type=Backbones2026.04 | 61.7 | — | — | 8.4 | 90 | |
| FIMA-QDetector=ViT-Base, Type=Backbones2026.04 | 61.1 | — | — | 26.4 | 92 | |
| FIMA-QDetector=Swin-Base, Type=Backbones2026.04 | 60.9 | — | — | 26.8 | 92 | |
| FIMA-QDetector=DeiT-Base, Type=Backbones2026.04 | 60.3 | — | — | 26.4 | 92 | |
| HOLaVenue=ICCV’25, Technology Category=Low-Rank Factorization, Backbone=DeiT-Base2026.04 | 59.2 | — | — | — | — | |
| X-PrunerVenue=CVPR’23, Technology Category=Model Pruning, Backbone=Swin-Base2026.04 | 59.1 | — | — | — | — | |
| AdalogDetector=ViT-Base, Type=Backbones2026.04 | 58.8 | — | — | 29.7 | 91 | |
| BRECQDetector=DeiT-Base, Type=Backbones2026.04 | 58.6 | — | — | 75.9 | 77 | |
| AdalogDetector=ViT-Small, Type=Backbones2026.04 | 58.2 | — | — | 10.1 | 88 | |
| FIMA-QDetector=LSDA, Type=CVPR’242026.04 | 58.1 | — | — | 22.8 | 91 | |
| BRECQDetector=ViT-Base, Type=Backbones2026.04 | 57.2 | — | — | 79.2 | 76 | |
| HOLaVenue=ICCV’25, Technology Category=Low-Rank Factorization, Backbone=ViT-Base2026.04 | 57 | — | — | — | — | |
| AdalogDetector=DeiT-Base, Type=Backbones2026.04 | 56.9 | — | — | 33 | 90 | |
| FIMA-QDetector=NPR, Type=CVPR’242026.04 | 56.7 | — | — | 17.8 | 90 | |
| One-bitVenue=NIPS’24, Technology Category=Low-Rank Factorization, Backbone=DeiT-Base2026.04 | 56.6 | — | — | — | — | |
| BRECQDetector=ViT-Small, Type=Backbones2026.04 | 56.4 | — | — | 21 | 75 | |
| AdalogDetector=LSDA, Type=CVPR’242026.04 | 56.3 | — | — | 27.9 | 88.9 | |
| BRECQDetector=Swin-Small, Type=Backbones2026.04 | 55.9 | — | — | 43.5 | 77 | |
| AdalogDetector=Swin-Base, Type=Backbones2026.04 | 55.8 | — | — | 33.5 | 90 | |
| X-PrunerVenue=CVPR’23, Technology Category=Model Pruning, Backbone=ViT-Base2026.04 | 55.5 | — | — | — | — | |
| HOLaVenue=ICCV’25, Technology Category=Low-Rank Factorization, Backbone=Swin-Base2026.04 | 55.4 | — | — | — | — | |
| AdalogDetector=NPR, Type=CVPR’242026.04 | 55.2 | — | — | 21.4 | 88 | |
| AdalogDetector=DeiT-Tiny, Type=Backbones2026.04 | 54.5 | — | — | 2.8 | 88.2 | |
| FIMA-QDetector=DeiT-Small, Type=Backbones2026.04 | 54.1 | — | — | 7.6 | 91 | |
| AdalogDetector=Swin-Small, Type=Backbones2026.04 | 53.7 | — | — | 20.8 | 88.9 | |
| FIMA-QDetector=DeiT-Tiny, Type=Backbones2026.04 | 53.2 | — | — | 2.4 | 90 | |
| One-bitVenue=NIPS’24, Technology Category=Low-Rank Factorization, Backbone=ViT-Base2026.04 | 53.2 | — | — | — | — | |
| AdalogDetector=DeiT-Small, Type=Backbones2026.04 | 52.8 | — | — | 9.3 | 88.9 | |
| BRECQDetector=LSDA, Type=CVPR’242026.04 | 52.8 | — | — | 58.2 | 77 | |
| One-bitVenue=NIPS’24, Technology Category=Low-Rank Factorization, Backbone=Swin-Base2026.04 | 52.5 | — | — | — | — | |
| BRECQDetector=DeiT-Tiny, Type=Backbones2026.04 | 52.1 | — | — | 5.9 | 75 | |
| GAC-FASVenue=CVPR’24, Technology Category=Neural Architecture Search, Backbone=DeiT-Base2026.04 | 52.1 | — | — | — | — | |
| X-PrunerVenue=CVPR’23, Technology Category=Model Pruning, Backbone=DeiT-Base2026.04 | 52.1 | — | — | — | — | |
| BRECQDetector=Swin-Base, Type=Backbones2026.04 | 51.5 | — | — | 73.7 | 78 | |
| GAC-FASVenue=CVPR’24, Technology Category=Neural Architecture Search, Backbone=ViT-Base2026.04 | 51.3 | — | — | — | — | |
| FIMA-QDetector=Wavelet, Type=WACV’252026.04 | 51.2 | — | — | 49.2 | 92 | |
| GAC-FASVenue=CVPR’24, Technology Category=Neural Architecture Search, Backbone=Swin-Base2026.04 | 50.6 | — | — | — | — | |
| BRECQDetector=NPR, Type=CVPR’242026.04 | 50.4 | — | — | 42.7 | 76 | |
| BRECQDetector=DeiT-Small, Type=Backbones2026.04 | 50.3 | — | — | 20.2 | 76 | |
| FIMA-QDetector=UCF, Type=ICCV’232026.04 | 50.1 | — | — | 10.1 | 89 | |
| ADDVenue=AAAI’22, Technology Category=Knowledge Distillation, Backbone=DeiT-Base2026.04 | 49.5 | — | — | — | — | |
| ADDVenue=AAAI’22, Technology Category=Knowledge Distillation, Backbone=Swin-Base2026.04 | 49 | — | — | — | — | |
| AdalogDetector=UCF, Type=ICCV’232026.04 | 48.2 | — | — | 11.9 | 87 | |
| AdalogDetector=Wavelet, Type=WACV’252026.04 | 47.8 | — | — | 61.5 | 90 | |
| ADDVenue=AAAI’22, Technology Category=Knowledge Distillation, Backbone=ViT-Base2026.04 | 46.6 | — | — | — | — | |
| BRECQDetector=Wavelet, Type=WACV’252026.04 | 43.9 | — | — | 135.3 | 78 | |
| BRECQDetector=UCF, Type=ICCV’232026.04 | 43.7 | — | — | 22.9 | 75 | |
| CORETraining dataset=FF++ (C23)2026.02 | — | — | 94.09 | — | — | |
| Face X-rayTraining dataset=FF++ (C23)2026.02 | — | — | 95.4 | — | — | |
| IIDTraining dataset=FF++ (C23)2026.02 | — | — | 93.92 | — | — | |
| MSBA-CLIPTraining dataset=FF++ (C23)2026.02 | — | — | 97.19 | — | — | |
| UCFTraining dataset=FF++ (C23)2026.02 | — | — | 93.1 | — | — |