Sample-level anomaly detection on AeBAD-S
85.6AUROC (Same)MMR
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| MMROptimizer=AdamW, Learning Rate=0.001, Training Epochs=200, Batch Size=162023.04 | 85.6 | 84.4 | 88.8 | 79.9 | 84.7 | |
| Reverse DistillationBackbone=WideResNet502023.04 | 82.4 | 84.3 | 85.5 | 71.9 | 81 | |
| PatchCoreBackbone=WideResNet502023.04 | 75.2 | 74.1 | 74.6 | 60.1 | 71 | |
| NSA2023.04 | 66.5 | 48.8 | 55.5 | 55.9 | 56.7 | |
| DRAEM2023.04 | 64 | 62.1 | 61.6 | 62.3 | 62.5 | |
| InTraImplementation=unofficial2023.04 | 39.8 | 46.1 | 44.7 | 46.3 | 44.2 | |
| RIADImplementation=unofficial2023.04 | 38.6 | 41.6 | 46.8 | 33 | 40 |