论文部分内容阅读
颜色是彩色图像最重要的视觉特征之一，在基于内容的图像检索（ＣＢＩＲ）系统中，都将颜色信息作为重要内容参与匹配和检索．针对图像中起主要视觉作用的是图像的主色这一问题，提出一种基于聚类分析的提取和表示图像主色的方法，给出一种用于聚类算法的停止准则．和等量量化方法相比，用本方法提取的颜色信息，具有特征维数低、颜色表示准的优点．
Color is one of the most important visual features of color image. In content-based image retrieval (CBIR) system, color information is taken as an important content to match and retrieve. In view of the problem that the main visual effect of the image is the dominant color of the image, a method of extracting and representing the dominant color of the image based on the clustering analysis is proposed, and a stopping criterion for the clustering algorithm is given. Compared with the equivalent quantification method, the color information extracted by this method has the advantages of low characteristic dimension and color representation.