Anomaly Detection on VisA 20% training contamination (test) benchmark leaderboard · SOTA2 Research

Anomaly Detection on VisA 20% training contamination (test)

0.723capsulesADL

Evaluation Results

Method
ADL2026.01	0.723	0.88	0.871	0.913	0.728	0.554	0.913	0.877	0.783	0.791	0.792	0.857	0.807
Multi-Cue Anomaly Detection framework2026.01	0.701	0.967	0.952	0.962	0.945	0.722	0.95	0.961	0.801	0.84	0.762	0.926	0.874
DestSeg2026.01	0.628	0.9	0.825	0.829	0.731	0.605	0.81	0.901	0.829	0.761	0.868	0.945	0.803
LOE2026.01	0.538	0.811	0.919	0.951	0.599	0.431	0.738	0.803	0.778	0.739	0.713	0.911	0.744
PatchCore2026.01	0.481	0.735	0.725	0.541	0.663	0.504	0.695	0.594	0.721	0.695	0.644	0.869	0.656
DRÆM2026.01	0.411	0.664	0.25	0.711	0.825	0.681	0.558	0.942	0.119	0.129	0.455	0.778	0.544
PaDiM2026.01	0.187	0.725	0.662	0.623	0.553	0.113	0.566	0.692	0.649	0.567	0.537	0.779	0.554