CFA

Reported on 6 benchmarks across 3 tasks · 2 papers

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· 2022-06-09
99.3
best: 99.9 (GLASS)
CFA: Coupled-hypersphere-based Feature Adaptation for Target-Oriented Anomaly Localization arXiv:2206.04325
Anomaly DetectiononMVTec AD
Segmentation AUROC· 2022-06-09
98.2
best: 99.7 (WeakREST-Block)
CFA: Coupled-hypersphere-based Feature Adaptation for Target-Oriented Anomaly Localization arXiv:2206.04325
Anomaly DetectiononVisA
Detection AUROC· 2022-06-09
92
best: 99.8 (UniNet)
CFA: Coupled-hypersphere-based Feature Adaptation for Target-Oriented Anomaly Localization arXiv:2206.04325
Anomaly DetectiononVisA
Segmentation AUPRO (until 30% FPR)· 2022-06-09
55.1
best: 96.1 (AnomalyDINO-S (full-shot))
CFA: Coupled-hypersphere-based Feature Adaptation for Target-Oriented Anomaly Localization arXiv:2206.04325
3DonKITTI Eigen split
absolute relative error· 2018-03-21
0.096
best: 0.029 (SPIDepth)
Monocular Depth Estimation by Learning from Heterogeneous Datasets arXiv:1803.08018

Computer Vision1 result

Depth EstimationonKITTI Eigen split
absolute relative error· 2018-03-21
0.096
best: 0.029 (SPIDepth)
Monocular Depth Estimation by Learning from Heterogeneous Datasets arXiv:1803.08018