基于改进颜色聚合向量与贡献度聚类的图像检索算法  

Image Retrieval Algorithm Based on Improved Color Coherence Vectors and Contribution to Clustering

作　　者：张永库[1] 李云峰[2] 孙劲光[1] 

机构地区：[1]辽宁工程技术大学电子与信息工程学院,葫芦岛125105 [2]辽宁工程技术大学研究生学院,葫芦岛125105

出　　处：《计算机科学》2015年第2期311-315,共5页Computer Science

基　　金：国家科技支撑计划(2013bah12f01)资助

摘　　要：为了提高图像检索的速度和准确率,通过分析各种聚类算法在图像检索中的缺点,提出了一种新的划分聚类的图像检索方法。首先,在对HSV模型非均匀量化的基础上,利用改进的颜色聚合向量方法提取图像的颜色特征;然后找到符合条件的特征向量作为初始聚类中心,利用分散度与贡献度进行聚类并建立特征索引库;最后根据查询图像的相似度进行检索和排序。实验结果表明,所提算法的查准率和查全率比其它算法均有较大提高。In order to improve the speed and accuracy of image retrieval,the drawbacks of image retrieval based on a variety of clustering algorithms were analyzed,and a new partition clustering method for image retrieval was presented in this paper.First,based on the asymmetrical quantization of the color in HSV model,color coherence vectors are introduced as the color feature.Secondly,qualified feature vectors are found as the initial cluster centers,and it clusteres based on the dispersion and the contribution,establishes image feature index library.Finally,it obtains the retrieval and reordered results by the similarity with the retrieval image.By comparing with other algorithms,it is demonstrated that the percentage of precision and recall of proposed algorithm are improved greatly.

关 键 词：HSV模型 颜色聚合向量 分散度 贡献度 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]