AI-generated image detection on FakeForm Poster
94.5AccuracyCoDA
Evaluation results
| Method | Links | ||
|---|---|---|---|
| CoDABackbone=ResNet2026.05 | 94.5 | 98.6 | |
| FatFormerBackbone=ViT2026.05 | 93.1 | 97.8 | |
| AIGI-HolmesBackbone=ViT2026.05 | 89.3 | 89.2 | |
| Humans2026.05 | 84.8 | 92.3 | |
| UFD2026.05 | 78.3 | 86.1 | |
| NPRBackbone=ResNet2026.05 | 63.1 | 79.5 | |
| AIDEBackbone=ResNeXt2026.05 | 63.1 | 68.7 | |
| SAFEBackbone=ResNet2026.05 | 62.3 | 70 | |
| FreqNetBackbone=ResNet2026.05 | 55.4 | 47.2 | |
| CNNDetBackbone=Xception2026.05 | 53.5 | 45 |