基于位平面理论和奇异值分解的鲁棒零水印算法  

Robust zero watermarking algorithm based on bit plane theory and singular value decomposition

作　　者：曲长波[1] 王东峰[1] 

机构地区：[1]辽宁工程技术大学软件学院,辽宁葫芦岛125105

出　　处：《计算机应用》2014年第12期3462-3465,3506,共5页journal of Computer Applications

摘　　要：针对灰度图像空域算法中水印鲁棒性较差的问题,提出了一种快速、鲁棒性强的零水印算法,并将该算法用于以数字图像为载体的信息隐藏,以实现水印信息的提取和认证。首先利用位平面(BP)理论将图像解析为不同的位平面层次,构造无权值的位平面矩阵,结合各位平面非零值个数生成图像的特征矩阵;然后对特征矩阵分块操作,利用奇异值分解(SVD)生成块最大奇异值矩阵,并通过对生成矩阵二维混沌加密得到注册零水印信息。仿真实验表明,所提算法具有较强的鲁棒性,在同类算法中对抗椒盐噪声攻击的性能提高了6%,对抗常见混合攻击的性能提高了12%。In view of the watermark robustness in information hiding algorithm of spatial domain, a zero watermarking algorithm which is fast and robust was proposed. And the algorithm was used in information hiding which was based on digital image in order to realize the watermark information extraction and certification. Firstly, the Bit Plane （BP） theory was used to analyze bit planes at different levels, set up the bit plane matrix structure which has no value, combine the numbers of non- zero values in bit planes to generate eigen matrix. Then, the eigen matrix was partitioned, using the singular value decomposition to the largest block singular value matrix was generated, and zero watermarking information was obtained by the matrix two-dimensional chaotic encryption registration. Simulation results show that, the proposed algorithm has high robustness against attacks, improved by 6% to salt and pepper noise attack than similar algorithms, and to common mixed attacks up to 12%.

关 键 词：位平面理论 零水印算法 旋转攻击 奇异值分解 混沌加密 

分 类 号：TP309.2[自动化与计算机技术—计算机系统结构]