Multimodal feature cooperative refinement for few-shot anomaly detection

Luo Xu, Delong Han, Gang Li, Mingle Zhou, Jin Wan, Min Li

Abstract: Highlights•A novel framework based on vision-language models for few-shot anomaly detection is proposed.•Image semantics is crucial for text prompts refinement.•Frequency and spatial features are complementary for few-shot anomaly detection.•The proposed method achieves superior performance compared to competitive methods.

External IDs:dblp:journals/aei/XuHLZWL25