基于内容的图像检索技术综述  

A survey on state-of-the-art techniques in content-based image retrieval

作　　者：刘颖[1] 范九伦[1] 

机构地区：[1]西安邮电学院通信与信息工程学院,陕西西安710121

出　　处：《西安邮电学院学报》2012年第2期1-8,共8页Journal of Xi'an Institute of Posts and Telecommunications

基　　金：陕西省"百人计划"支持项目

摘　　要：基于内容的图像检索技术(CBIR)是近些年来图像检索领域的研究热点,其发展经历了最初的以图像数字特征为索引,到引入图像语义学习技术使得CBIR更贴近人类语义以方便用户查询,再到如今融合多种图像信息以进一步提高检索效率。本文在对已有文献进行综述的基础上,结合近几年的研究成果,讨论CBIR的最新技术动态,分析CBIR的发展趋势并指出几个未来的研究方向。From retrieval by visual image features to introducing image semantics for more convenient searching,and then to the fusion of multiple features for more accurate retrieval,content-based image retrieval（CBIR） has been a hot research topic for.over three decades.Based on the earlier literature survey done by the authors and a study on the state-of-art techniques in CBIR with high-level semantics,this paper explains the improvement that has been made in this area since then and the problems that are yet to be solved.In addition,a few future research topics are identified.

关 键 词：图像检索 图像语义学习 图像特征提取 信息融合 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]