Anomaly Detection on MVTec-AD (Img AUROC, AP)
98.4Img AUROCUniDG-RFT
Evaluation results
| Method | Links | ||
|---|---|---|---|
| UniDG-RFTLearning Paradigm=Training-Free Few-shot Image Insertion, Synthesis Strategy=RFT2026.04 | 98.4 | 99 | |
| UniDG-SFTLearning Paradigm=Training-Free Few-shot Image Insertion, Synthesis Strategy=SFT2026.04 | 98 | 98.6 | |
| AnoDiffLearning Paradigm=Few-shot Learning Anomaly Generation2026.04 | 97.9 | 98.9 | |
| AnoGenLearning Paradigm=Few-shot Learning Anomaly Generation2026.04 | 97 | 95 | |
| SeaSLearning Paradigm=Few-shot Learning Anomaly Generation2026.04 | 96.2 | 95.5 | |
| VMADSetting=Zero-shot, Evaluation=Cross-dataset2024.09 | 95.8 | 87.6 | |
| InsertAnyLearning Paradigm=Training-Free Few-shot Image Insertion2026.04 | 95.3 | 95.5 | |
| DualAnoDiffLearning Paradigm=Few-shot Learning Anomaly Generation2026.04 | 94.9 | 96.6 | |
| MyriadSetting=Zero-shot, Evaluation=Cross-dataset2024.09 | 93.2 | 88.2 | |
| WinClipSetting=Zero-shot, Evaluation=Cross-dataset2024.09 | 92.5 | 87.5 | |
| AnomalyGPTSetting=Zero-shot, Evaluation=Cross-dataset2024.09 | 92.1 | 85.1 | |
| VANDSetting=Zero-shot, Evaluation=Cross-dataset2024.09 | 91.8 | 86.8 | |
| AnomalyClipSetting=Zero-shot, Evaluation=Cross-dataset2024.09 | 91.5 | 83.4 | |
| BGADSetting=Zero-shot, Evaluation=Cross-dataset2024.09 | 90.1 | 85.6 | |
| LISASetting=Zero-shot, Evaluation=Cross-dataset2024.09 | 89.1 | 81.2 | |
| AnyDoorLearning Paradigm=Training-Free Few-shot Image Insertion2026.04 | 80.3 | 80.7 |