Cheqing Jin

Robust Zero-shot Anomaly Detection under Limited Auxiliary Anomaly Priors

Guanyu Lu, Fang Zhou, Cheqing Jin

Zero-shot anomaly detection aims to identify defects in arbitrary novel domains; however, existing models assume that the auxiliary data contains a rich diversity of anomalies, neglecting the far more complex and unpredictable variations in real-world target domains. This study i…Read more