ReConPatch WRN-50

Reported on 4 benchmarks across 1 task · 1 paper · 1 SOTA

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Methodology4 results

Anomaly DetectiononBTAD
Segmentation AUPRO· 2023-05-26
97.5
SOTA
ReConPatch : Contrastive Patch Representation Learning for Industrial Anomaly Detection arXiv:2305.16713
Anomaly DetectiononBTAD
Detection AUROC· 2023-05-26
95.8
best: 98.3 (CPR)
ReConPatch : Contrastive Patch Representation Learning for Industrial Anomaly Detection arXiv:2305.16713
Anomaly DetectiononMVTec AD
Detection AUROC· uses extra data· 2023-05-26
99.56
best: 99.9 (GLASS)
ReConPatch : Contrastive Patch Representation Learning for Industrial Anomaly Detection arXiv:2305.16713
Anomaly DetectiononMVTec AD
Segmentation AUROC· uses extra data· 2023-05-26
98.18
best: 99.7 (WeakREST-Block)
ReConPatch : Contrastive Patch Representation Learning for Industrial Anomaly Detection arXiv:2305.16713