基于图像内容和支持向量机的服装图像检索方法研究  

A Clothing Image Retrieval Method Based on Image Content and Support Vector Machine

作　　者：薛培培[1] 邬延辉[1] 

机构地区：[1]宁波大学,浙江宁波315211

出　　处：《移动通信》2016年第2期79-82,87,共5页Mobile Communications

摘　　要：基于内容的图像检索技术在电商领域的应用已逐渐成为研究热点。根据服装图像的特点,提出了一种基于颜色特征和尺度不变特征转换(SIFT)并进行特征融合的服装图像检索方法,在此基础上利用支持向量机加入相关反馈技术,来提高图像的检索准确率。实验结果表明,提出的方法可以提高图像的检索准确率。Content-based image retrieval technology has become a popular topic in the field of electronic commerce. According to characteristics of clothing image, a feature-fused clothing image retrieval method combined color feature with scale-invariant feature transform（SIFT） was proposed. Support vector machine was also used to add feedback to enhance the accuracy of image retrieval. Experimental results show that the proposed method can improve the accuracy of image retrieval.

关 键 词：图像检索 颜色直方图 SIFT特征 支持向量机 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]