AI Image Detection on StyleGAN3
90AccuracyPGC
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| PGCTraining Data=ProGAN (4 classes) and SDv1.42026.05 | 90 | — | — | — | 97.7 | |
| SAFETraining Data=ProGAN (4 classes) and SDv1.42026.05 | 89.7 | — | — | — | 97.6 | |
| AIDETraining Data=ProGAN (4 classes) and SDv1.42026.05 | 88.1 | — | — | — | 91.4 | |
| FreqNetTraining Data=ProGAN (4 classes) and SDv1.42026.05 | 83 | — | — | — | 92.4 | |
| UnivFDTraining Data=ProGAN (4 classes) and SDv1.42026.05 | 79.6 | — | — | — | 84.5 | |
| NPRTraining Data=ProGAN (4 classes) and SDv1.42026.05 | 78.4 | — | — | — | 91.7 | |
| LGradTraining Data=ProGAN (4 classes) and SDv1.42026.05 | 70.5 | — | — | — | 80.5 | |
| B-FreeTraining Data=ProGAN (4 classes) and SDv1.42026.05 | 62.5 | — | — | — | 79.3 | |
| DDATraining Data=ProGAN (4 classes) and SDv1.42026.05 | 58.3 | — | — | — | 76.2 | |
| CNNDetTraining Data=ProGAN (4 classes) and SDv1.42026.05 | 55.8 | — | — | — | 73.1 | |
| ViTModel Architecture=ViT, Training Dataset=Stable Diffusion2025.12 | 48.31 | 46.37 | 89.39 | 61.06 | — | |
| ResNet50Backbone=ResNet50, Training Data=Stable Diffusion2025.12 | 45.34 | 45.34 | 100 | 62.39 | — |