一种新的基于颜色和空间特征的图像检索方法  

A New Approach for Image Retrieval Based on Color and Spatial Features

作　　者：孙君顶[1] 张喜民[1] 崔江涛[1] 周利华[1] 

机构地区：[1]西安电子科技大学多媒体研究所,西安710071

出　　处：《计算机科学》2005年第6期158-160,184,共4页Computer Science

基　　金：十五国防科技(电子)预研项目(413160501)

摘　　要：提出了一种新的基于颜色和空间特征的图像检索方法。首先提取色调不变量作为图像的颜色特征,然后设计了图像状态矩阵来描述图像的形状信息和空间位置信息,在进行相似检索时,采用 Guassian模型对不同子特征间的距离进行了归一化处理。试验表明利用该算法提取的图像的颜色和空间特征在进行图像检索时效果显著。A new content-based color image retrieval method is proposed in this paper,in which both the color content and the spatial features have been taken into account. Based on the special disposal on the HSV color space,the image invariable,hue is selected as the color character,and then an improved accumulative histogram of the hue is calculated as the color feature. To attain the image's shape and spatial information,the image's H-,S-,and V- component are first transformed into status matrix. Then the state correlogram of the status matrix is calculated as the image's spa- tial distribution information. In matching the sindlarity of images,the Guassian model is used to normalize the differ- ent sub-characters distance. Experiments indicate that this method is great effective in boage's similarity retrieval,and has a higher retrieval-rate.

关 键 词：图像检索方法 空间特征 归一化处理 颜色特征 位置信息 形状信息 状态矩阵 相似检索 效果显著 不变量 和空间 提取 算法 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TN911.25[自动化与计算机技术—计算机科学与技术]