基于分形和HVS的小波域数字图像水印算法  

Watermark Algorithm in DWT Domain Based on Fractal Theory and HVS

作　　者：张波[1] 王宇欣[1] 李天颖[1] 杨树国[1] 

机构地区：[1]青岛科技大学数理学院,山东青岛266100

出　　处：《电脑与电信》2016年第7期5-8,21,共5页Computer & Telecommunication

基　　金：山东省高等学校科研计划项目;项目编号:J13LN34;山东省重点研发计划项目;项目编号:2015GGX101020;青岛市科技发展计划项目;项目编号:KJZD-13-27-JCH;2016年青岛科技大学大学生创新创业训练计划项目;项目编号:201606001

摘　　要：针对数字图像版权的保护,本文提出了一种基于分形的小波域数字图像水印算法。首先将载体图像进行Haar小波变换,将小波系数树分为同方向小波树,并在同方向小波树中构造域子树和值子树,通过分形匹配将域子树分为嵌入域子树和非嵌入域子树,然后在嵌入域子树中加入水印。在水印嵌入时,只嵌入部分水印信息来减少水印的嵌入量,同时运用人类视觉系统HVS模型控制水印嵌入的强度,保证其不可见的同时最大限度增强了其鲁棒性并实现了水印的盲提取。In order to protect the copyright of the image, this paper proposes a kind of watermark algorithm in DWT domain based on fractal theory and HVS. Firstly, the original image is transformed via DWT and the wavelet coefficients are divided into isotropic wavelet trees, then the domain subtrees and range subtrees are constructed in isotropic wavelet trees. It will divide domain subtrees into embedded domain subtrees and non- embedded domain subtrees by fractal theory matching,then the watermark is added in the embedded domain subtrees. We only embed part of information of the image and use the HVS module to control the capacity of the embedded watermark, which maximizes the robustness when the watermark is not visible and achieves the blind watermark extraction.

关 键 词：小波变换 分形 人类视觉系统 数字盲水印 鲁棒性 

分 类 号：TP309.7[自动化与计算机技术—计算机系统结构]