DMAD: Dual Memory Bank for Real-World Anomaly Detection

arXiv:2403.12362v1 Announce Type: cross
Abstract: Training a unified model is considered to be more suitable for practical industrial anomaly detection scenarios due to its generalization ability and storage efficiency. However, this multi-class setting, which exclusively uses normal data, overlooks the few but important accessible annotated anomalies in the real world. To address the challenge of real-world anomaly detection, we propose a new framework named Dual Memory bank enhanced representation learning for Anomaly Detection (DMAD). This framework handles both unsupervised and …

abstract anomaly anomaly detection arxiv bank class cs.cv cs.lg data detection efficiency however industrial memory normal practical storage training type unified model world