一种新的数字图像检索方法的研究  

Research on a New Digital Image Retrieval Algorithm

作　　者：安嵘[1] 任雪萍[1] 

机构地区：[1]杭州电子科技大学计算机学院,浙江杭州310018

出　　处：《杭州电子科技大学学报（自然科学版）》2008年第1期73-76,共4页Journal of Hangzhou Dianzi University：Natural Sciences

摘　　要：该文提出了一种改进的聚类算法来实现基于颜色的图像检索,并给出了基于分块主色的检索方法描述图像颜色特征。算法有效地解决了K-均值聚类算法初始聚类中心的选取问题。实验结果表明该算法与其他算法相比具有较好的准确性和检索效率,检索的效果和速度都有了提高。This article proposed one kind of improved cluster algorithm to realize image retrieval based on the color,and described the image color characteristic based on the piecemeal essential color.The algorithm has effectively solved the K-average value cluster algorithm initial cluster center selection problem. The experimental result indicated this algorithm compared with other algorithms has the good accuracy and the retrieval efficiency,the retrieval effect and the speed all had the enhancement.

关 键 词：聚类 图像检索 分块主色 

分 类 号：TP391.3[自动化与计算机技术—计算机应用技术]