基于Hessenberg分解的彩色图像盲水印算法  

Color blind image watermarking algorithm based on Hessenberg decomposition

作　　者：韩绍程 王蕊[2] 贾桂敏 

机构地区：[1]中国民航大学基础实验中心,天津300300 [2]中国民航大学理学学院,天津300300 [3]中国民航大学电子信息与自动化学院,天津300300

出　　处：《计算机工程与设计》2017年第12期3354-3360,共7页Computer Engineering and Design

基　　金：国家自然科学基金项目(61379102;61502498;61601467;U1433105;U1433120);中央高校基本业务费专项基金项目(3122014D030)

摘　　要：针对彩色图像的版权保护问题,提出一种基于分数阶混沌系统和Hessenberg分解的彩色图像水印算法。提取载体图像在YIQ彩色空间中的Y分量,对其进行Contourlet变换,将获得的低频子带进行不重叠分块,通过自适应量化每个子块Hessenberg分解后上三角矩阵中最大能量元素来嵌入水印。利用分数阶混沌系统的随机特性预处理水印信息,增强算法的安全性,水印提取不需要原始图像。实验结果表明,该算法对添加噪声、中值滤波、行列随机删除、区域剪切和亮度及对比度调整等攻击具有较好的鲁棒性。A color image watermarking algorithm was proposed based on fractional-order chaotic system and Hessenberg decomposition for the copyright protection of color images.The Y component of the original color image in YIQ space was performed with Contourlet transform.The low frequency sub-band was divided into non-overlapping blocks.The watermark bits were embedded into the maximum energy elements of the upper triangular matrix from the blocks experienced the Hessenberg decomposition by the adaptive quantization.To enhance the security of the algorithm,the watermark was preprocessed using random characteristics of the fractional-order chaotic system.The watermark was extracted without the original image.Experimental results show that the proposed algorithm is robust to many kinds of attacks,such as adding noise,median filter,row or column random deletion,area cutting,brightness and contrast adjustment and so on.

关 键 词：数字水印 彩色图像 分数阶混沌系统 Hessenberg分解 CONTOURLET变换 

分 类 号：TP309.7[自动化与计算机技术—计算机系统结构]