Anomaly Detection On Mvtec Loco Ad
Metrics
Avg. Detection AUROC
Detection AUROC (only logical)
Detection AUROC (only structural)
Segmentation AU-sPRO (until FPR 5%)
Results
Performance results of various models on this benchmark
Comparison Table
| Model Name | Avg. Detection AUROC | Detection AUROC (only logical) | Detection AUROC (only structural) | Segmentation AU-sPRO (until FPR 5%) |
|---|---|---|---|---|
| sub-image-anomaly-detection-with-deep-pyramid | 68.9 | 70.9 | 66.8 | 45.1 |
| ladmim-logical-anomaly-detection-with-masked | 86.0 | 83.1 | 90.3 | - |
| slsg-industrial-image-anomaly-detection-by | 90.3 | 89.6 | 91.4 | 67.3 |
| csad-unsupervised-component-segmentation-for | 95.3 | 96.7 | 94.0 | - |
| visual-anomaly-detection-via-dual-attention | 83.7 | 79.2 | 88.2 | 67.4 |
| towards-total-recall-in-industrial-anomaly | 80.3 | 75.8 | - | 39.7 |
| asymmetric-student-teacher-networks-for | - | 79.7 | 87.1 | 42.7 |
| set-features-for-fine-grained-anomaly | 86.8 | 88.9 | 84.7 | - |
| few-shot-part-segmentation-reveals | 94.9 | 98.1 | 91.6 | - |
| component-aware-anomaly-detection-framework | 89.8 | 90.1 | 89.4 | - |
| uninformed-students-student-teacher-anomaly | 77.3 | 66.4 | 88.3 | - |
| towards-total-recall-in-industrial-anomaly | 79.4 | 71.0 | 87.7 | 36.5 |
| dsr-a-dual-subspace-re-projection-network-for | 82.6 | 75.0 | 90.2 | 58.5 |
| learning-memory-guided-normality-for-anomaly | 65.1 | 60.0 | 70.2 | 33.9 |
| hard-nominal-example-aware-template-mutual | 88.1 | 83.2 | 92.9 | - |
| set-features-for-anomaly-detection | 88.3 | 91.2 | 85.5 | - |
| contextual-affinity-distillation-for-image | 84.0 | 81.2 | 86.9 | 73.0 |
| f-anogan-fast-unsupervised-anomaly-detection | 64.2 | 65.8 | 62.7 | 33.4 |
| puad-frustratingly-simple-method-for-robust-1 | 93.1 | 92.0 | 94.1 | - |
| set-features-for-anomaly-detection | 94.2 | 95.8 | 94.2 | - |
| generating-and-reweighting-dense-contrastive | 87.5 | - | - | - |
| beyond-dents-and-scratches-logical | 57.3 | 58.1 | 56.5 | 37.8 |
| learning-global-local-correspondence-with | 83.1 | 82.4 | 83.8 | 70.3 |
| component-aware-anomaly-detection-framework | 81.2 | 87.7 | 74.6 | - |
| musc-zero-shot-industrial-anomaly | 75.9 | 67.47 | 84.3 | 63.04 |
| beyond-dents-and-scratches-logical | 57.7 | 56.5 | 58.9 | 22.5 |
| fastflow-unsupervised-anomaly-detection-and | 79.2 | 75.5 | 82.9 | 56.8 |
| template-guided-hierarchical-feature | 86.0 | 85.2 | 86.7 | 74.1 |
| auto-encoding-variational-bayes | 54.3 | 53.8 | 54.8 | 38.2 |
| draem-a-discriminatively-trained | 73.6 | 72.8 | 74.4 | 42.6 |
| component-aware-anomaly-detection-framework | 88.2 | 87.5 | 88.8 | - |
| sam-lad-segment-anything-model-meets-zero | 90.7 | - | - | 83.2 |
| efficientad-accurate-visual-anomaly-detection | 90.7 | 86.8 | 94.7 | 79.8 |
| anomaly-detection-via-reverse-distillation | 78.7 | 69.4 | 88.0 | 63.7 |
| simplenet-a-simple-network-for-image-anomaly | 77.6 | 71.5 | 83.7 | 36.3 |
| puad-frustratingly-simple-method-for-robust-1 | 94.4 | 93.7 | 95.9 | - |
| beyond-dents-and-scratches-logical | 83.3 | 86.0 | 80.6 | 70.1 |
| component-aware-anomaly-detection-framework | 90.1 | 89.4 | 90.9 | - |
| efficientad-accurate-visual-anomaly-detection | 90.0 | 85.8 | 94.1 | 77.8 |
| component-aware-anomaly-detection-framework | 87.9 | 85.9 | 89.9 | - |