Anomaly Detection on MVTec-AD 15-class (I-AUC variants)
98.4I-AUC (Average)DeCoFlow
Evaluation results
| Method | Links | ||
|---|---|---|---|
| DeCoFlowContinual Learning Setting=1x1 CIL, Training Protocol=Incremental2026.06 | 98.4 | 0 | |
| Joint_PatchCoreContinual Learning Setting=1x1 CIL, Training Protocol=Joint2026.06 | 97.8 | — | |
| Joint_CADICContinual Learning Setting=1x1 CIL, Training Protocol=Joint2026.06 | 97.8 | — | |
| CADICContinual Learning Setting=1x1 CIL, Training Protocol=Incremental2026.06 | 97.2 | 1.1 | |
| ReplayCADContinual Learning Setting=1x1 CIL, Training Protocol=Incremental2026.06 | 94.8 | 4.5 | |
| UCADContinual Learning Setting=1x1 CIL, Training Protocol=Incremental2026.06 | 93 | 1 | |
| DNEContinual Learning Setting=1x1 CIL, Training Protocol=Incremental2026.06 | 87 | 11.6 | |
| IUFContinual Learning Setting=1x1 CIL, Training Protocol=Incremental2026.06 | 76.2 | 6.7 | |
| CDADContinual Learning Setting=1x1 CIL, Training Protocol=Incremental2026.06 | 74.3 | 20.1 | |
| FT_SimpleNetContinual Learning Setting=1x1 CIL, Training Protocol=Fine-tuning2026.06 | 70.8 | 21.1 | |
| FT_CFAContinual Learning Setting=1x1 CIL, Training Protocol=Fine-tuning2026.06 | 62.3 | 36.1 | |
| FT_PatchCoreContinual Learning Setting=1x1 CIL, Training Protocol=Fine-tuning2026.06 | 60.2 | 38.3 | |
| FT_RD4ADContinual Learning Setting=1x1 CIL, Training Protocol=Fine-tuning2026.06 | 59.6 | 39.3 |