AI-generated Image Detection on FakeForm Hi-MAR
91AccuracyAIGI-Holmes
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| AIGI-HolmesBackbone=ViT, FPS=2.2, Params (M)=> 70002026.05 | 91 | 92 | |
| AIDEBackbone=ResNeXt, FPS=18.9, FLOPs=225.69, Params (M)=893.542026.05 | 90.2 | 91.5 | |
| CoDABackbone=ResNet, FPS=125.2, FLOPs=2.37, Params (M)=1.482026.05 | 89.5 | 91.2 | |
| SAFEBackbone=ResNet, FPS=180.4, FLOPs=2.29, Params (M)=1.442026.05 | 87 | 90.5 | |
| FreqNetBackbone=ResNet, FPS=157.2, FLOPs=1.97, Params (M)=1.852026.05 | 83.5 | 88.6 | |
| FatFormerBackbone=ViT, FPS=13.3, FLOPs=216.83, Params (M)=492.592026.05 | 82.5 | 88.8 | |
| UFD2026.05 | 78.2 | 85.8 | |
| Humans2026.05 | 74 | 84 | |
| CNNDetBackbone=Xception, FPS=26.2, FLOPs=16.57, Params (M)=23.512026.05 | 70.8 | 80.2 | |
| NPRBackbone=ResNet, FPS=200.6, FLOPs=1.75, Params (M)=1.442026.05 | 65.8 | 73.5 |