基于广义直图的图像检索方法  

An Efficient Content-based Retrieval Method for Images

作　　者：王禹[1] 李弼程[1] 郭志刚[1] 魏俊[1] 

机构地区：[1]郑州信息工程大学信息科学系,郑州450002

出　　处：《计算机工程与应用》2002年第20期81-83,共3页Computer Engineering and Applications

摘　　要：文中提出了一种基于广义直方图的图像内容检索方法。首先,论文把图像和它的平滑图像组合起来,形成一个二元组,称之为“广义图像”,广义图像的直方图称为广义直方图;其次,把图像颜色空间聚类算法推广到广义图像;最后,利用传统直方图方法的检索原理,根据广义直方图对广义图像进行检索,从而实现对原图像的检索。实验结果表明,新方法非常有效,与传统直方图方法相比,较大地提高了输出精度。A new method of image indexing based on Generalized color histogram is presented here.Firstly,the paper combines the original image with its smoothed image to form a binary set.lt call this binary set as a 'generalized image', and the histogram of the generalized image is a generalized histogram.Secondly, the algorithm of color cluster is extended to generalized image.Finally, It can index an image by its generalized histogram with the basic principle of the classic histogram method.Experiment results show that the new method is very efficient, and the precision of the indexing output is improved largely compared with the classic histogram method.

关 键 词：广义直图 图像检索 颜色直方图 广义图像 彩色图像 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]