WinCLIP+ (2-shot)

Reported on 5 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.

Methodology5 results

Anomaly DetectiononMVTec AD
Detection AUROC· 2023-03-26
94.4
best: 99.9 (GLASS)
WinCLIP: Zero-/Few-Shot Anomaly Classification and Segmentation arXiv:2303.14814
Anomaly DetectiononMVTec AD
Segmentation AUPRO· 2023-03-26
88.4
best: 98.4 (WeakREST-Block)
WinCLIP: Zero-/Few-Shot Anomaly Classification and Segmentation arXiv:2303.14814
Anomaly DetectiononVisA
Detection AUROC· 2023-03-26
84.6
best: 99.8 (UniNet)
WinCLIP: Zero-/Few-Shot Anomaly Classification and Segmentation arXiv:2303.14814
Anomaly DetectiononVisA
Segmentation AUPRO· 2023-03-26
86.2
best: 96 (DiffusionAD)
WinCLIP: Zero-/Few-Shot Anomaly Classification and Segmentation arXiv:2303.14814
Anomaly DetectiononVisA
Segmentation AUPRO (until 30% FPR)· 2023-03-26
86.2
best: 96.1 (AnomalyDINO-S (full-shot))
WinCLIP: Zero-/Few-Shot Anomaly Classification and Segmentation arXiv:2303.14814