基于颜色空间特性的图像检索  

Image Retrieval Based on Color-Spatial Features

作　　者：李延龙[1] 李太君[1] 罗其朝[1] 

机构地区：[1]海南大学信息科学技术学院,海南海口570228

出　　处：《海南大学学报（自然科学版）》2013年第4期344-348,共5页Natural Science Journal of Hainan University

基　　金：海南省科技兴海专项资金项目(XH201311)

摘　　要：提出一种能同时利用图像的颜色特征和颜色空间分布信息进行图像检索的方法,首先提取出最大的3块颜色区域及各自的最小包围矩形(MBR),然后利用改进的子块技术,确定目标区域的位置,从而得到融合颜色与位置信息的索引,在此基础上实现图像的检索.实验结果证明了该方法的有效性.In our report, an image retrieval method that can use spatial distribution information of color and color features of the image at the same time was proposed. Firstly, three dominant color regions, which are the biggest in an image, together with its respective Minimum Bounding Rectangle （MBR） were extracted. Secondly, the improved sub-block technology was used to determine the location of the target area and obtain composite color- spatial index to realize the image retrieval. The results showed the effectiveness of the nronosed methnd.

关 键 词：图像检索 颜色特征 颜色空间分布 子块技术 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]