Dinomaly ViT-B (model-unified multi-class)

Reported on 4 benchmarks across 1 task · 1 paper

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

Methodology4 results

Anomaly DetectiononMVTec AD
Detection AUROC· uses extra data· 2024-05-23
99.6
best: 99.9 (GLASS)
Dinomaly: The Less Is More Philosophy in Multi-Class Unsupervised Anomaly Detection arXiv:2405.14325
Anomaly DetectiononMVTec AD
Segmentation AP· uses extra data· 2024-05-23
69.29
best: 87.6 (WeakREST-Block)
Dinomaly: The Less Is More Philosophy in Multi-Class Unsupervised Anomaly Detection arXiv:2405.14325
Anomaly DetectiononMVTec AD
Segmentation AUPRO· uses extra data· 2024-05-23
94.79
best: 98.4 (WeakREST-Block)
Dinomaly: The Less Is More Philosophy in Multi-Class Unsupervised Anomaly Detection arXiv:2405.14325
Anomaly DetectiononMVTec AD
Segmentation AUROC· uses extra data· 2024-05-23
98.35
best: 99.7 (WeakREST-Block)
Dinomaly: The Less Is More Philosophy in Multi-Class Unsupervised Anomaly Detection arXiv:2405.14325