基于颜色聚类的JPEG图像索引与检索  

Indexing and Retrieval of JPEG Images Based on Color Clusters

作　　者：石跃祥[1] 易唐唐[1] 谭勇[1] 

机构地区：[1]湘潭大学信息工程学院,湖南湘潭411105

出　　处：《湘潭大学自然科学学报》2008年第4期126-129,共4页Natural Science Journal of Xiangtan University

基　　金：博士科研启动基金资助项目(06QDZ23);湖南省自科基金资助项目(07JJ6115)

摘　　要：针对JPEG压缩图像,根据感知相似性,提出一种基于颜色聚类带相关反馈的快速压缩域索引和检索方法.该方法直接在YCbCr空间上进行,可提高图像特征提取的速度、节省存储空间.实验结果证明了本方法的有效性.A fast compressed domain image indexing and retrieval method based on color clusters and relevance feedback was presented＇ which aimed at JPEG images and kept to color peceptual similarity. As the method was directly operating in the YCbCr space, the speed of extracting image features was improved improved andstorage space saved. The experimental results demonstrate the effectiveness of the proposed method.

关 键 词：感知相似性 颜色索引 压缩域 图像检索 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]