基于灰色关联度的自适应图像多小波域密写  

Steganography of self-adaptive images in multiwavelet domain based on grey correlation degree

作　　者：武月红[1] 张毓森[1] 朱欣刚[2] 

机构地区：[1]解放军理工大学指挥自动化学院,江苏南京210007 [2]中国联通辽宁分公司,辽宁沈阳110003

出　　处：《解放军理工大学学报（自然科学版）》2010年第2期184-189,共6页Journal of PLA University of Science and Technology(Natural Science Edition)

基　　金：国家自然科学基金资助项目(60073049)

摘　　要：多小波变换解决密写问题的缺点是多比特嵌入容易破坏系数间的相关性,为了解决这个问题,首先运用灰色关联分析计算邻域内多小波树组的中心树序列与外围树序列的相关程度,根据相关性的大小自适应地控制系数的可替换比特数,并在LL、LH和HL频带系数进行比特替换。对多幅图像满嵌后的隐藏容量和含密图像PSNR值的实验结果表明,该算法作为变换域的隐藏方法容量较大,隐蔽性很好,实现了在满足隐蔽性的前提下进一步提高容量的目的。另外,抗攻击检验结果表明,从受到攻击的含密图像中提取的秘密图像的质量较好,具有一定的稳健性。The shortcoming of muhiwavelet steganography is that multi-bits embedding can easily destroythe correlation of coefficients. To solve the problem, grey relational analysis was firstly used to calculate the correlation degree between the central tree sequence and the peripheral trees sequence of trees group in the adjoining region, the number of bits which could be replaced was self-adaptively controlled, and then the bits replacements to LL, LH and HI. subband coefficients were done. Experiments of capacity and PSNR for many fully embedded stegano images indicate that the steganography has good imperceptibility and large capacity as an information hiding method, and realizes the goal of enhancing the capacity under the precondition of imperceptibility. In addition, the anti-attack tests show that the secret images recovered from the attacked stegano images have high quality and the steganography has some solidity.

关 键 词：数字密写 灰色关联分析 视觉特性 多小波变换 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]