基于规范类正交矩阵的信息隐藏算法  

Information hiding algorithm based on normal similar-orthogonal matrices

作　　者：唐燕[1,2] 闾国年[1] 殷奎喜[3] 

机构地区：[1]南京师范大学虚拟地理环境教育部重点实验室,南京210046 [2]南京师范大学泰州学院,泰州225300 [3]南京师范大学物理科学与技术学院,南京210046

出　　处：《东南大学学报（自然科学版）》2013年第1期45-49,共5页Journal of Southeast University：Natural Science Edition

基　　金：国家高技术研究发展计划(863计划)资助项目(2008105GZ30031)

摘　　要：提出了一种规范类正交矩阵,该矩阵中任意2行的互相关系数为定值,且矩阵列数不受限制.首先,将隐秘数据经过规范类正交矩阵编码,嵌入载体图像的2×2分块离散余弦变换系数矩阵的同频分量内,以提高信息隐藏算法的安全性.其次,基于参量数据的贪心算法,实现规范类正交矩阵的快速构造.最后,利用邻域信息的相似性,实现隐秘数据的正确提取.实验数据的分析结果表明,信息嵌入量、组块数、互相关参数、检测错误率等指标之间存在相互制约的关系.模拟结果表明,所提出的信息隐藏算法具有图像失真小、嵌入容量大、鲁棒性强的特点.A kind of normal similar-orthogonal （NS） matrices is defined. The cross-correlation coefficient of any two rows in the NS matrices is a constant and the number of columns has no bounds. First, the secret information is coded by a NS matrix and embedded into the same frequency components of 2 × 2 discrete cosine transform coefficients of the original image, which can improve the security of the information hiding algorithm. Then, by introducing a greedy algorithm based on parameters, the fast generator of NS matrices is realized. Finally, by using image neighboring rela- tivity, the secret information can be totally extracted. The analysis results of empirical data show that there are trade-off relations among the embedded information number, the group number, the cross- correlation parameter and the detection error probability. The simulation results demonstrate that the proposed method has low image distortion, high embedding capacity and high robust.

关 键 词：规范类正交矩阵 信息隐藏 离散余弦变换 误码率 

分 类 号：TH918.3[机械工程]