AI-generated image detection on FakeForm Photojournalism
0.873AccuracyCoDA
Evaluation results
| Method | Links | ||
|---|---|---|---|
| CoDABackbone=ResNet2026.05 | 0.873 | 0.94 | |
| FatFormerBackbone=ViT2026.05 | 0.834 | 0.915 | |
| AIGI-HolmesBackbone=ViT2026.05 | 0.803 | 0.867 | |
| Humans2026.05 | 0.793 | 0.879 | |
| UFD2026.05 | 0.732 | 0.771 | |
| NPRBackbone=ResNet2026.05 | 0.705 | 0.834 | |
| SAFEBackbone=ResNet2026.05 | 0.635 | 0.658 | |
| AIDEBackbone=ResNeXt2026.05 | 0.62 | 0.796 | |
| FreqNetBackbone=ResNet2026.05 | 0.581 | 0.623 | |
| CNNDetBackbone=Xception2026.05 | 0.554 | 0.601 |