FiLo: Zero-Shot Anomaly Detection by Fine-Grained Description and High-Quality Localization

arXiv:2404.13671v1 Announce Type: cross
Abstract: Zero-shot anomaly detection (ZSAD) methods entail detecting anomalies directly without access to any known normal or abnormal samples within the target item categories. Existing approaches typically rely on the robust generalization capabilities of multimodal pretrained models, computing similarities between manually crafted textual features representing "normal" or "abnormal" semantics and image features to detect anomalies and localize anomalous patches. However, the generic descriptions of "abnormal" often fail to precisely match diverse types of anomalies across different …

abstract access anomaly anomaly detection arxiv capabilities computing cs.cv cs.lg detection features fine-grained localization multimodal normal pretrained models quality robust samples textual type zero-shot