采用两次模运算的无直方图异常安全密写法  

Steganography Using Dual-Modulo Arithmetic without Introducing Histogram Abnormality

作　　者：王威娜[1] 周振宇[1] 王朔中[1] 

机构地区：[1]上海大学通信与信息工程学院,上海200072

出　　处：《上海大学学报（自然科学版）》2006年第4期331-337,共7页Journal of Shanghai University:Natural Science Edition

基　　金：国家自然科学基金资助项目(60372090);上海市科委基础研究重点项目(04JC14037)

摘　　要：提出一种以像素对为基础的数字图像密写方法,该方法同时对像素灰度值和像素对的差值进行模运算(modulo gray-value modulo difference,简称MGMD),在每一对像素中可嵌入的比特数取决于像素对差值的大小.基于待嵌入的数据和两次模运算的结果修改像素对,并维持其差值不变.MGMD密写不会引起像素直方图和像素差值直方图的异常,不仅避免了一些基于像素对的密写法如PVD和SM的安全漏洞,还能抵御针对常规LSB密写的RS分析和χ2分析.该方法的安全性得到实验的验证.A new steganographic embedding technique based on pixel pairs is described. The proposed method involves two modulo operations on the difference between pixels in each pixel pair and the gray-value of a pixel, respectively, hence the name modulo gray-value modulo difference（MGMD）. The number of bits to be embedded into each pair is determined by the pixel difference. The secret data and the results of two modulo operations are combined and used to modify the host pixel pair such that the gray level difference is kept unchanged. MGMD embedding causes no histogram abnormality as often encountered in other pixel pair-based techniques such as PVD and SM. It can also resist the popular RS and X^2 analyses. Security of the method is confirmed in experiments on a large number of images.

关 键 词：安全密写 模运算 像素差值 直方图 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]