AI-generated image detection on StyleGAN generator (test)
99.42AccuracyFatFormer
Evaluation results
| Method | Links | ||
|---|---|---|---|
| FatFormerTraining Dataset=ProGAN2026.01 | 99.42 | 99.99 | |
| NPRTraining Dataset=ProGAN2026.01 | 99.25 | 100 | |
| Ours (ViT-L/14)Training Dataset=ProGAN, Backbone=ViT-L/142026.01 | 94.59 | 99.97 | |
| SAFETraining Dataset=ProGAN2026.01 | 93.24 | 99.96 | |
| Ours (ViT-B/32)Training Dataset=ProGAN, Backbone=ViT-B/322026.01 | 91.66 | 98.56 | |
| LNPTraining Dataset=ProGAN2026.01 | 91.2 | 97.37 | |
| LGradTraining Dataset=ProGAN2026.01 | 89.63 | 98.03 | |
| CNNSpotTraining Dataset=ProGAN2026.01 | 84.64 | 99.36 | |
| GramNetTraining Dataset=ProGAN2026.01 | 83.59 | 94.49 | |
| UnivFDTraining Dataset=ProGAN2026.01 | 80.4 | 97.48 | |
| FreDectTraining Dataset=ProGAN2026.01 | 78.02 | 88.98 |