基于内容的医学图像检索方法是近年来计算机视觉领域的研究热点,已经广泛应用于计算机辅助诊断的研究中。概述了基于内容的医学图像检索方法的研究进展及意义,介绍了当前主流的医学图像检索算法及其优缺点,旨在引导研究人员快速了解本领域的研究内容。医学图像检索的研究主要分为特征提取和相似性度量两部分。从传统特征提取及近年来兴起的基于深度学习的特征提取入手来介绍医学图像的特征提取方式;而相似性度量部分则详细列举了马氏距离度量、词汇树以及哈希算法。最后概述了医学图像检索领域的相关反馈技术及当前常用的图像检索系统,并讨论了医学图像检索未来可能的研究方向及相关难点。

Content-based medical image retrieval method is a research hotspot in the field of computer vision in recent years, and has been widely used in the research of computer-aided diagnosis. This paper summarizes the research progress and significance of content-based medical image retrieval methods, introduces the current mainstream medical image retrieval algorithms and their advantages and disadvantages, and aims to guide researchers to quickly understand the research content in this field. The research of medical image retrieval is mainly divided into two parts: feature extraction and similarity measurement. This paper introduces the feature extraction method of medical images starting with the extraction of traditional features and the feature extraction based on deep learning emerging in recent years. The similarity measure part enumerates the Mahalanobis distance metric, vocabulary tree, and hash algorithm. Finally, the related feedback technology in the field of medical image retrieval and the commonly used image retrieval system are summarized. The possible research directions and related difficulties in medical image retrieval are discussed.