AI-generated image detection on FakeForm CT Scan
98.6AccuracyFatFormer
Evaluation results
| Method | Links | ||
|---|---|---|---|
| FatFormerBackbone=ViT2026.05 | 98.6 | 99.8 | |
| NPRBackbone=ResNet2026.05 | 95.5 | 98.5 | |
| FreqNetBackbone=ResNet2026.05 | 94.8 | 99.2 | |
| AIGI-HolmesBackbone=ViT2026.05 | 94.8 | 96.3 | |
| CoDABackbone=ResNet2026.05 | 92.3 | 94.8 | |
| SAFEBackbone=ResNet2026.05 | 89.2 | 96.5 | |
| AIDEBackbone=ResNeXt2026.05 | 83.4 | 93.6 | |
| UFD2026.05 | 76.3 | 89.5 | |
| Humans2026.05 | 75.4 | 88.3 | |
| CNNDetBackbone=Xception2026.05 | 75.4 | 88.3 |