图像预测误差域的动态补偿安全隐写算法  

Dynamic Compensation Secure Steganographic Algorithm in Prediction-error Domain

作　　者：张湛[1] 刘光杰[1] 王俊文[1] 戴跃伟[1] 王执铨[1] 

机构地区：[1]南京理工大学自动化学院,江苏南京210094

出　　处：《南京理工大学学报》2010年第2期187-192,共6页Journal of Nanjing University of Science and Technology

基　　金：江苏省自然科学基金(BK2008403)

摘　　要：为了保证载密图像嵌入大量信息后的高阶统计安全性,该文将图像马尔可夫链模型推广至图像预测误差域,根据动态补偿的原理提出运用于图像预测误差域的二阶统计保持隐写算法。该算法根据隐写对图像预测误差马尔可夫链模型统计分布的改变情况,指导嵌入函数,运用后嵌入的隐秘信息动态补偿前期嵌入对图像二阶统计特性的改变,从而在不损失嵌入容量的情况下提高隐写的统计安全性。实验表明,该隐写算法在较大容量的数据嵌入过程中,能很好保持图像统计特性,提高隐写安全性。In order to ensure the high-order statistical security of stego-images into which large amount of information is embedded,the image Markov chain model is extended to the prediction-error domain of image.Based on the dynamic compensation theory,a novel steganographic algorithm preserving second-order statistical properties in prediction-error domain is proposed.According to the changes of cover-image prediction-error Markov chain statistical distribution,the algorithm guides embedded functions to make use of post-embedding secret information to compensate dynamically the second-order statistical changes of cover-image caused by pre-embedding.The statistical security of stego-image is improved without loss of the steganographic capacity.The experimental results show： when large-capacity information is embedded,the proposed algorithm has high ability to preserve the second-order statistical properties in prediction-error domains of cover images and to improve the steganographic security.

关 键 词：信息隐藏 隐写 动态补偿 马尔可夫链 统计二阶保持 预测误差域 

分 类 号：TP309[自动化与计算机技术—计算机系统结构]