Anomaly Detection on VisA supervised benchmark 12 different classes
97.7Image-level AUROCDPDiff-AD
Evaluation results
| Method | Links | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| DPDiff-AD2026.05 | 97.7 | — | — | 98.5 | 95.5 | 99.1 | 49.2 | 51.5 | 94.5 | 83.7 | |
| ReContrast2026.05 | 95.5 | — | — | 96.4 | 92 | 98.5 | 47.9 | 50.6 | 91.9 | 81.8 | |
| MambaAD2026.05 | 94.3 | — | — | 94.5 | 89.4 | 98.5 | 39.4 | 44 | 91 | 78.7 | |
| HVQ-Trans2026.05 | 93.2 | — | — | 92.8 | 87.6 | 98.7 | 35 | 39.6 | 86.3 | 76.2 | |
| RD4AD2026.05 | 92.4 | — | — | 92.4 | 89.6 | 98.1 | 38 | 42.6 | 91.8 | 77.8 | |
| UniAD2026.05 | 91.4 | — | — | 93.3 | 87.5 | 98.5 | 35.3 | 40.2 | 89 | 76.5 | |
| ViTAD2026.05 | 90.5 | — | — | 91.7 | 86.3 | 98.2 | 36.6 | 41.1 | 85.1 | 75.6 | |
| DeSTSeg2026.05 | 88.9 | — | — | 89 | 85.2 | 96.1 | 39.6 | 43.4 | 67.4 | 72.8 | |
| SimpleNet2026.05 | 87.2 | — | — | 87 | 81.8 | 96.8 | 34.7 | 37.8 | 81.4 | 72.4 | |
| DiAD2026.05 | 86.8 | — | — | 88.3 | 85.1 | 96 | 26.1 | 33 | 75.2 | 70.1 | |
| All AD + SegADparadigm=Supervised Anomaly Detection, base_detectors=RD4AD and EfficientAD2024.05 | — | 98.4 | 7.8 | — | — | — | — | — | — | — | |
| DevNetparadigm=Supervised Anomaly Detection2024.05 | — | 89.3 | 44 | — | — | — | — | — | — | — | |
| DRAparadigm=Supervised Anomaly Detection2024.05 | — | 88.9 | 42.4 | — | — | — | — | — | — | — | |
| EfficientADparadigm=One-class Anomaly Detection2024.05 | — | 97.9 | 8.3 | — | — | — | — | — | — | — | |
| EfficientAD + SegADparadigm=Supervised Anomaly Detection, base_detector=EfficientAD2024.05 | — | 98.3 | 8.1 | — | — | — | — | — | — | — | |
| RD4ADparadigm=One-class Anomaly Detection2024.05 | — | 94.7 | 24.5 | — | — | — | — | — | — | — | |
| RD4AD + SegADparadigm=Supervised Anomaly Detection, base_detector=RD4AD2024.05 | — | 95.3 | 19.3 | — | — | — | — | — | — | — |