LECLIP: Boosting Zero-Shot Anomaly Detection With Local Enhanced CLIP

doi:10.1109/TIM.2025.3571124

	LECLIP: Boosting Zero-Shot Anomaly Detection With Local Enhanced CLIP
	Liu, Yuyao 1; Li, Qingyong 2; Wang, Zhehong 1; Kato, Jien 3; Zhang, Jie 4; Wang, Wen 1
	2025
发表期刊	IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
ISSN	0018-9456
卷号	74 页码:11
摘要	Zero-shot anomaly detection (ZSAD) is a critical task that detects anomalies without any training samples from the target application, which is crucial for applications in diverse fields such as industrial quality control and medical imaging analysis. Recent advances have seen the application of contrastive language-image pretraining (CLIP) in ZSAD, exploiting its robust visual-linguistic alignment and zero-shot learning capabilities. However, CLIP is primarily designed for natural image classification, emphasizing global visual embeddings, while anomaly detection (AD) requires a more accurate representation of anomalous regions and more precise local visual embeddings. To overcome these limitations, this article proposes the local enhanced CLIP (LECLIP) framework for ZSAD. LECLIP incorporates a local alignment (LA) module that divides images into blocks and aligns them with learnable text embeddings, ensuring precise relevance expression. Furthermore, a training-free echo-attention (EA) is proposed to complement the traditional QKV attention, enabling the model to capture both global and local image details effectively, thus providing a more accurate and detailed image representation. Experimental results show that LECLIP achieves superior performance on 15 challenging datasets, including six industrial datasets and nine medical datasets. Code is available at https://github.com/lyy70/LECLIP
关键词	Visualization Anomaly detection Training Accuracy Image representation Feature extraction Biomedical imaging Semantics Attention mechanisms Artificial intelligence Echo-attention (EA) local alignment (LA) module local enhanced CLIP (LECLIP) zero-shot anomaly detection (ZSAD)
DOI	10.1109/TIM.2025.3571124
收录类别	SCI
语种	英语
资助项目	Fundamental Research Funds for the Central Universities[2024QYBS026] ; Fundamental Research Funds for the Central Universities[2023JBZY037] ; Fundamental Research Funds for the Central Universities[2022JBMC055] ; Fundamental Research Funds for the Central Universities[2022JBQY007] ; Beijing Natural Science Foundation[L231019] ; National Natural Science Foundation of China[62276019] ; National Natural Science Foundation of China[62306028] ; Langfang Research and Development Projects[2023011003B] ; Shenzhen Science and Technology Program Project[KJZD20240903102742055]
WOS研究方向	Engineering ; Instruments & Instrumentation
WOS类目	Engineering, Electrical & Electronic ; Instruments & Instrumentation
WOS记录号	WOS:001504194800024
出版者	IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
引用统计
文献类型	期刊论文
条目标识符	http://119.78.100.204/handle/2XEOYT63/42348
专题	中国科学院计算技术研究所期刊论文_英文
通讯作者	Wang, Wen
作者单位	1.Beijing Jiaotong Univ, Key Lab Big Data & Artificial Intelligence Transpo, Minist Educ, Beijing 100044, Peoples R China 2.Beijing Jiaotong Univ, Frontiers Sci Ctr Smart High Speed Railway Syst, Beijing 100044, Peoples R China 3.Kochi Univ Technol, Sch Data & Innovat, Kochi 7828502, Japan 4.Chinese Acad Sci, Inst Comp Technol, Beijing 100864, Peoples R China
推荐引用方式 GB/T 7714	Liu, Yuyao,Li, Qingyong,Wang, Zhehong,et al. LECLIP: Boosting Zero-Shot Anomaly Detection With Local Enhanced CLIP[J]. IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT,2025,74:11.
APA	Liu, Yuyao,Li, Qingyong,Wang, Zhehong,Kato, Jien,Zhang, Jie,&Wang, Wen.(2025).LECLIP: Boosting Zero-Shot Anomaly Detection With Local Enhanced CLIP.IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT,74,11.
MLA	Liu, Yuyao,et al."LECLIP: Boosting Zero-Shot Anomaly Detection With Local Enhanced CLIP".IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT 74(2025):11.