AI-generated image detection on BigGAN generator (test)
99.68AccuracyOurs (ViT-L/14)
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Ours (ViT-L/14)Training Dataset=ProGAN, Backbone=ViT-L/142026.01 | 99.68 | 99.99 | |
| I2P2026.04 | 97.64 | 99.59 | |
| FatFormerTraining Dataset=ProGAN2026.01 | 96.83 | 99.82 | |
| FatFormerVenue=CVPR 20242026.04 | 96.67 | 99.47 | |
| UnivFDTraining Dataset=ProGAN2026.01 | 94.5 | 99.27 | |
| ForgeLensVenue=ICCV 20252026.04 | 93.8 | 99.07 | |
| Ours (ViT-B/32)Training Dataset=ProGAN, Backbone=ViT-B/322026.01 | 92.68 | 98.94 | |
| FerretNetVenue=NIPS 20252026.04 | 91.32 | 99.58 | |
| FreqNetVenue=AAAI 20242026.04 | 90.58 | 94.81 | |
| SAFETraining Dataset=ProGAN2026.01 | 90.25 | 95.71 | |
| RINEVenue=ECCV 20242026.04 | 82.8 | 95.46 | |
| FreDectTraining Dataset=ProGAN2026.01 | 81.97 | 93.62 | |
| NPRVenue=CVPR 20242026.04 | 81.84 | 88.64 | |
| LNPTraining Dataset=ProGAN2026.01 | 81.8 | 90.97 | |
| LGradTraining Dataset=ProGAN2026.01 | 81.73 | 89.08 | |
| UniFDVenue=CVPR 20232026.04 | 81.21 | 96.13 | |
| EffortVenue=ICML 20252026.04 | 77.6 | 98.38 | |
| NPRTraining Dataset=ProGAN2026.01 | 76.85 | 83.1 | |
| CNNSpotTraining Dataset=ProGAN2026.01 | 69.3 | 81.89 | |
| GramNetTraining Dataset=ProGAN2026.01 | 67.55 | 62.34 | |
| DIREVenue=CVPR 20232026.04 | 50.66 | 54.58 | |
| LGardVenue=CVPR 20232026.04 | 40.69 | 39.36 |