基于代表色特征的图像检索

A CONTENT BASED IMAGE RETRIVAL METHODUSING REPRESENTATIVE COLORS FEATORES

基于内容的图像检索 ( CBIR,Content- based Image Retrieval)在多媒体检索中得到越来越多的重视 .本文提出了一种基于图像颜色特征的图像检索方法 .在经过变形的 HSL颜色模型空间 ,通过适当的聚类分析算法提取出能够代表图像颜色特征的代表色集 ,并引入代表色的分布特征和纹理特征 ,作为图像的特征矢量进行匹配检索 .在试验中本方法取得了较好的试验结果 .

Content-based Image Retrieval plays an important role in multimedia retrieval field. In this paper, a CBIR method using picture's color features is discussed. In the deformed HSL model space, we use a suitable cluster algorithm to extract color information of an image, then the distributive features and texture features of representative color are used in image mach process. The experiment give good results.