基于图像距离测度的置乱效果评价法  

Scrambling quality evaluation method based on image distance measure

作　　者：吴成茂[1,2] 田小平[1,2] 谭铁牛[2] 

机构地区：[1]西安邮电学院电子与信息工程系,陕西西安710121 [2]中国科学院自动化研究所模式识别国家重点实验室,北京100080

出　　处：《计算机工程与设计》2009年第21期4891-4894,共4页Computer Engineering and Design

基　　金：中国科学院自动化研究所模式识别国家重点实验室开放课题基金项目(07-31-3)

摘　　要：提出了一种修改的图像距离进行图像置乱效果评价的方法。针对传统欧氏距离不能真实刻画置乱图像间差异程度的不足,考虑到图像中任意位置灰度信息与其周围邻域灰度存在紧密相关,将一种具有黎曼度量特性的图像距离进行修改并用来构造一种描述图像置乱效果好坏的评价函数。通过大量的实验得出,该评价方法是能够保持主客观评价的一致性。A kind of new image scrambling quality evaluation method is presented by image distance measure modified. Aiming at the shortcoming of Euclidean distance that is not factually capable of measuring scrambled image differentia degree. Considering that the arbitrary position image gray information is compactly associated with its neighbor image gray information, a kind of new modified image distance is proposed, then the evaluation function of describing image scrambling quality good or bad is constructed by new modified image distance with Riemann measure properties. The conclusion of which objective evaluation method can be consistent with subjective evaluation method is drawn by many experimental results.

关 键 词：图像置乱 置乱度 欧氏距离 图像距离 空间几何关系 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]