基于二维直方图信息熵的图像检索方法  

Image Retrieval Based on Two-dimensional Histogram Information Entropy

作　　者：洪清启[1] 王备战[1] 史亮[1] 董槐林[1] 陈兵[1] 

机构地区：[1]厦门大学软件学院,福建厦门361005

出　　处：《广西师范大学学报（自然科学版）》2007年第4期265-268,共4页Journal of Guangxi Normal University:Natural Science Edition

基　　金：985二期信息创新平台项目资助课题(0000-X07204)

摘　　要：针对传统颜色直方图完全丢失颜色的空间分布信息的缺点,提出了基于二维直方图信息熵的图像检索方法。该方法不仅利用了图像象素的颜色统计信息,还用到了相邻像素之间的相关性及颜色在图像内的空间分布特征,充分考虑到了图像的颜色统计、空间信息等特征。首先提取出图像的二维直方图特征;接着,为降低特征的维度,计算出二维直方图相应的信息熵并用其进行相似性度量。实验表明该方法实现简单,而且能够较准确地描述出图像的颜色统计、空间分布等信息,从而有效地提高了检索的精度。The color histogram based on image retrieval method is simple and efficient, however, it has completely lost the information of color spatial distribution. This paper proposes a new retrieval method based on two-dimensional histogram. The method considers not only the star info of the color of image pixels but also the relativity between border upon pixels and the information of color spatial distribution in the picture. Firstly, it picks-up the feature of two dimensional histogram, and then in order to reduce the dimension of the character, it figures out the entropy of the two-dimensional histogram and processes the comparability measurement. The experiment results show that the method is more accurate to describe the features of the stat info of the color and the information of color spatial distribution and improves the image retrieval precision.

关 键 词：基于内容的图像检索 颜色直方图 二维直方图 信息熵 

分 类 号：O391[理学—工程力学]