图像信息隐藏不可感知性指标研究  

Study on Imperceptibility Index of Information Hidden Based on Image

作　　者：谢建全[1,2] 谢勍[2] 黄大足[1,2] 阳春华[1] 

机构地区：[1]中南大学信息科学与工程学院,湖南长沙410083 [2]湖南财政经济学院信息管理系,湖南长沙410205

出　　处：《小型微型计算机系统》2011年第5期953-957,共5页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(60574030)资助

摘　　要：分析目前衡量信息隐藏算法不可感知性指标的不足之处,根据人类视觉特性,提出一种新的有原始载体图像作对照时的视觉不可感知性评价指标,该指标综合考虑了载体图像的背景照度、背景纹理复杂性、空间位置及其方向性对视觉掩蔽效应的影响,试验证明该指标能比峰值信噪比及加权峰值信噪比更好地衡量隐藏信息的不可感知性.它不但能评价灰度图像,而且能评价二值图像,解决了灰度图像与二值图像不能使用同一评价指标的问题.At first,the deficiency of imperceptibility indices of nowadays information hiding algorithm is analyzed.Then,a new evaluation indices system of good invisible characters with the comparative original host image is proposed according to human vision system（HVS）.This indices system considers comprehensively the background illumination,the complexity of background texture,the spatial position of host image and the Influence of its direction on mask effect of human vision system.Experiments show that this indices system can be used to evaluate the imperceptibility of information hidden better than power signal-to-noise ratio（PSNR） and weighted PSNR（WPSNR）.It can be used to evaluate not only gray image but also binary image,and the problem,which gray image and binary image can not make use of the same evaluation indices.

关 键 词：信息隐藏 不可感知性 性能评价 峰值信噪比 视觉失真感知函数 

分 类 号：TP309[自动化与计算机技术—计算机系统结构]