Towards Total Recall in Industrial Anomaly Detection

Benchmark Model Rank Results
3d-anomaly-detection-on-anomaly-shapenet10PatchCore (FPFH)#2O-AUROC: 0.884P-AUROC: 0.923
3d-anomaly-detection-on-anomaly-shapenet10PatchCore (PointMAE)#4O-AUROC: 0.574P-AUROC: 0.645
3d-anomaly-detection-on-real-3d-adPatchCore (FPFH+Raw)#7Mean Performance of P. and O.: 0.687Point AUROC: 0.692
3d-anomaly-detection-on-real-3d-adPatchCore (PointMAE)#10Mean Performance of P. and O.: 0.614Point AUROC: 0.634
3d-anomaly-detection-on-real-3d-adPatchCore (FPFH)#12Mean Performance of P. and O.: 0.5925Point AUROC: 0.592
anomaly-classification-on-goodsadPatchCore-100%#1AUPR: 86.1AUROC: 85.5
anomaly-classification-on-goodsadPatchCore-1%#2AUPR: 83.3AUROC: 81.4
anomaly-detection-on-aebad-sPatchCore#4Detection AUROC: 71.0Segmentation AUPRO: 87.8
anomaly-detection-on-aebad-vPatchCore#3Detection AUROC: 70.7
anomaly-detection-on-mpddPatchCore#10Detection AUROC: 82.12Segmentation AUROC: 95.66
anomaly-detection-on-mvtec-adPatchCore Large#17Detection AUROC: 99.6Segmentation AUPRO: 93.5Segmentation AUROC: 98.2
anomaly-detection-on-mvtec-adPatchCore#34Detection AUROC: 99.2Segmentation AUROC: 98.4
anomaly-detection-on-mvtec-adPatchCore(16shot)#74Detection AUROC: 95.4
anomaly-detection-on-mvtec-loco-adPatchCore#18Avg. Detection AUROC: 80.3Detection AUROC (only logical): 75.8
anomaly-detection-on-mvtec-loco-adPatchCore Ensemble#19Avg. Detection AUROC: 79.4Detection AUROC (only logical): 71.0