Structural Teacher-Student Normality Learning for Multi-Class Anomaly Detection and Localization

arXiv:2402.17091v1 Announce Type: new
Abstract: Visual anomaly detection is a challenging open-set task aimed at identifying unknown anomalous patterns while modeling normal data. The knowledge distillation paradigm has shown remarkable performance in one-class anomaly detection by leveraging teacher-student network feature comparisons. However, extending this paradigm to multi-class anomaly detection introduces novel scalability challenges. In this study, we address the significant performance degradation observed in previous teacher-student models when applied to multi-class anomaly detection, which we identify as resulting from cross-class …

abstract anomaly anomaly detection arxiv class cs.cv data detection distillation feature knowledge localization modeling network normal normality paradigm patterns performance set type visual