Video-level Deepfake Detection on DFD
0.996AUCKFD
Evaluation results
| Method | Links | |
|---|---|---|
| KFDDFD (Dedicated Deepfake Detector)=true, MLLM (Multimodal Large Language Model)=true, Training dataset=FF++ c232026.01 | 0.996 | |
| M2F2-detDFD (Dedicated Deepfake Detector)=true, MLLM (Multimodal Large Language Model)=true, Training dataset=FF++ c232026.01 | 0.977 | |
| VLAForgeVenue=-2026.03 | 0.972 | |
| EffortDFD (Dedicated Deepfake Detector)=true, MLLM (Multimodal Large Language Model)=false, Training dataset=FF++ c232026.01 | 0.965 | |
| RLSBIDFD (Dedicated Deepfake Detector)=false, MLLM (Multimodal Large Language Model)=true, Training dataset=FF++ c232026.01 | 0.965 | |
| Yan et al.Venue=CVPR’252026.03 | 0.965 | |
| EffortVenue=ICML’252026.03 | 0.965 | |
| ForAdaVenue=CVPR’252026.03 | 0.965 | |
| SAMVenue=CVPR’242026.03 | 0.961 | |
| X2-DFDDFD (Dedicated Deepfake Detector)=true, MLLM (Multimodal Large Language Model)=true, Training dataset=FF++ c232026.01 | 0.957 | |
| LSDAVenue=CVPR’242026.03 | 0.956 | |
| CDFADFD (Dedicated Deepfake Detector)=true, MLLM (Multimodal Large Language Model)=false, Training dataset=FF++ c232026.01 | 0.954 | |
| CDFAVenue=ECCV’242026.03 | 0.954 | |
| RepDFDVenue=AAAI’252026.03 | 0.951 | |
| IIDVenue=CVPR’232026.03 | 0.939 | |
| ProDetDFD (Dedicated Deepfake Detector)=true, MLLM (Multimodal Large Language Model)=false, Training dataset=FF++ c232026.01 | 0.901 | |
| ProDetVenue=Neurips’242026.03 | 0.901 | |
| RECCEDFD (Dedicated Deepfake Detector)=true, MLLM (Multimodal Large Language Model)=false, Training dataset=FF++ c232026.01 | 0.891 | |
| SBIVenue=CVPR’222026.03 | 0.882 | |
| LSDADFD (Dedicated Deepfake Detector)=true, MLLM (Multimodal Large Language Model)=false, Training dataset=FF++ c232026.01 | 0.881 | |
| SFDGVenue=CVPR’232026.03 | 0.88 | |
| UCFDFD (Dedicated Deepfake Detector)=true, MLLM (Multimodal Large Language Model)=false, Training dataset=FF++ c232026.01 | 0.867 | |
| TALLVenue=ICCV’232026.03 | 0.833 | |
| SBIDFD (Dedicated Deepfake Detector)=true, MLLM (Multimodal Large Language Model)=false, Training dataset=FF++ c232026.01 | 0.827 | |
| LAA-NETVenue=CVPR’242026.03 | 0.8 |