基于内容的图像检索技术在服装检索中的应用  

The Application of Content Based Image Retrieval Technology in Clothing Retrieval System

作　　者：王海龙[1] 杜俊俐[1] 郭清宇[1] 

机构地区：[1]中原工学院计算机学院,河南郑州450007

出　　处：《计算技术与自动化》2009年第2期88-91,共4页Computing Technology and Automation

基　　金：河南省科技攻关项目(0624220102)

摘　　要：随着服装电子商务的日益普及,基于图像的服装检索需求日渐迫切。颜色、花型、款式是服装最本质的三大特征,基于这三大特征进行研究,用代表色表达服装的颜色,利用灰度共生矩阵刻画服装的花型,提取外轮廓特征表达服装的款式,提出相应的匹配算法,实现对服装图像的全方位检索。实验表明,达到理想的检索效果。With the increasing popularity of clothing e- commerce, the requirements of Image- based clothing retrieval are increasingly urgent. The eolor, flower type and style of clothing are the three easential eharaeteristies. Under the research of three characteristics, the paper presents the method of extraction, expression and matching to implement clothing image retrieval comprehensively,using representative color, gray level eo -occurrenee matrix extraction, the characteristics of contour to express clothing image. Experiments show the method can achieve desired effeet.

关 键 词：服装图像检索 颜色特征 花型特征 款式特征 灰度共生矩阵 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]