AI-generated image detection on ADM generator (test)
84.39AccuracyNPR
Evaluation results
| Method | Links | ||
|---|---|---|---|
| NPRTraining Dataset=ProGAN2026.01 | 84.39 | 99.41 | |
| SAFETraining Dataset=ProGAN2026.01 | 79.41 | 96.75 | |
| FreDectTraining Dataset=ProGAN2026.01 | 69.8 | 74.97 | |
| FatFormerTraining Dataset=ProGAN2026.01 | 55.95 | 82.54 | |
| Ours (ViT-L/14)Training Dataset=ProGAN, Backbone=ViT-L/142026.01 | 54.11 | 90.69 | |
| UnivFDTraining Dataset=ProGAN2026.01 | 53.48 | 81.15 | |
| Ours (ViT-B/32)Training Dataset=ProGAN, Backbone=ViT-B/322026.01 | 52.38 | 89.09 | |
| CNNSpotTraining Dataset=ProGAN2026.01 | 50.2 | 61.66 | |
| GramNetTraining Dataset=ProGAN2026.01 | 50.02 | 56.65 | |
| LNPTraining Dataset=ProGAN2026.01 | 49.65 | 39.02 | |
| LGradTraining Dataset=ProGAN2026.01 | 47.65 | 34.95 |