摘要
基于内容的图像检索(CBIR)是对传统信息检索领域的扩展.它采用图像视觉内容的相似性判别进行查询.CBIR涉及到很多科学领域的课题.本文则仅主要综述CBIR技术中的相似性度量方法,索引方式,以及检索性能的评价.最后,分析了该领域现存的问题、最新研究动态及发展方向.
Content Based Image Retrieval（CBIR）is an extension to the traditional information retrieval. It adopts visual content rather than keyword to judge image similarity. CBIR systems involve a broad area of scientific research. This paper only gives a comprehensive survey of similarity measurement, feature organization and access methods, and performance evaluation in CBIR. Finally, the survey analyses the open issues, research trends and promising directions in this area.
出处
《小型微型计算机系统》
CSCD
北大核心
2007年第2期356-360,共5页
Journal of Chinese Computer Systems
关键词
图像检索
相似性度量
高维索引
性能评价
image retrieval
similarity measure
high-dimensional indexing
performance evaluation
