基于主色调的图像布局特征提取及相似性匹配算法  

Spatial layout features extraction based on primary color and new algorithm for image features matching

作　　者：黄德才[1] 钟艳平[1] 胡嘉[1] 

机构地区：[1]浙江工业大学信息工程学院,浙江杭州310032

出　　处：《浙江工业大学学报》2007年第1期46-50,共5页Journal of Zhejiang University of Technology

基　　金：浙江省自然科学基金资助项目(Y105118)

摘　　要：在基于内容的图像检索中,图像特征的提取和特征匹配是图像检索的核心.为了克服普通颜色特征不考虑颜色空间分布的缺点,提出了一种基于主色调的空间布局特征提取方法,并在此基础上提出了基于布局特征和颜色特征的图像相似性匹配算法-双特征匹配算法.仿真实验表明,该算法对一些布局特征明显的图像类型有很好的检索效果.In Content-Based Image Retrieval（CBIR） area, features extraction and matching are the key technique for image retrieval. In order to overcome the shortage of common color feature without regarding to spatial layout features, an approach to extract spatial layout features is presented based on pictures＇ primary color, and a new algorithm（double features algorithm） is presented too for image features matching by combined the spatial layout feature and color feature. The numerical experiments show that the new algorithm can achieve much better performance for a class of image with typical spatial layout feature.

关 键 词：图像检索 主色调 布局特征 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]