AI-generated image detection on Wukong generator (test)
96.45AccuracyFatFormer
Evaluation results
| Method | Links | ||
|---|---|---|---|
| FatFormerTraining Dataset=ProGAN2026.01 | 96.45 | 96.89 | |
| SAFETraining Dataset=ProGAN2026.01 | 91.95 | 96.65 | |
| Ours (ViT-L/14)Training Dataset=ProGAN, Backbone=ViT-L/142026.01 | 86.56 | 97.84 | |
| Ours (ViT-B/32)Training Dataset=ProGAN, Backbone=ViT-B/322026.01 | 84.35 | 95.49 | |
| NPRTraining Dataset=ProGAN2026.01 | 78.9 | 85.5 | |
| GramNetTraining Dataset=ProGAN2026.01 | 62.65 | 59.61 | |
| LGradTraining Dataset=ProGAN2026.01 | 60.55 | 62.31 | |
| LNPTraining Dataset=ProGAN2026.01 | 58.3 | 64.9 | |
| UnivFDTraining Dataset=ProGAN2026.01 | 54.35 | 80.36 | |
| CNNSpotTraining Dataset=ProGAN2026.01 | 50.05 | 55.02 | |
| FreDectTraining Dataset=ProGAN2026.01 | 41.7 | 42.04 |