Anomaly Detection on MVTec AD 15% training contamination (test)

0.991Carpet AUROC (MVTec AD)Multi-Cue Anomaly Detection framework

Evaluation results

Method
Multi-Cue Anomaly Detection framework2026.01	0.991	0.999	0.923	0.93	0.999	0.989	0.875	0.863	0.945	0.837	0.958	0.982	0.995	0.997	0.927	0.947
ADL2026.01	0.931	1	0.767	0.859	0.953	0.998	0.805	0.824	0.928	0.841	0.886	0.971	0.999	0.978	0.885	0.908
LOE2026.01	0.925	0.981	0.829	0.782	0.875	0.913	0.839	0.901	0.991	0.545	0.738	0.712	0.984	0.951	0.437	0.827
DestSeg2026.01	0.828	1	0.561	0.937	0.998	0.917	0.678	0.882	0.941	0.852	0.919	0.913	1	0.978	0.995	0.893
PaDiM2026.01	0.817	0.738	0.495	0.758	0.701	0.55	0.215	0.656	0.533	0.445	0.616	0.588	0.941	0.346	0.334	0.582
DRÆM2026.01	0.758	0.958	0.637	0.891	0.99	0.645	0.297	0.902	0.929	0.805	0.874	0.938	1	0.961	0.917	0.833
PatchCore2026.01	0.681	0.831	0.797	0.782	0.791	0.642	0.827	0.742	0.755	0.739	0.739	0.755	0.721	0.621	0.605	0.735