图像属性的自适应盲水印算法  

Adaptive blind watermarking algorithm based on image characteristics

作　　者：李璠[1] 邓承志[1] 

机构地区：[1]南昌工程学院计算机科学与技术系,江西南昌330099

出　　处：《南昌工程学院学报》2008年第6期12-15,共4页Journal of Nanchang Institute of Technology

基　　金：江西省教育厅科技项目(GJJ09366)

摘　　要：提出一种基于图像属性和小波包变换的盲数字水印算法.算法利用小波包变换对图像特征的稀疏表示特性,将图像分成不同内容属性的区域:平滑区、边缘区和纹理区;找出各区域块中视觉重要信息的位置,并将水印嵌入其中;水印嵌入强度取决于图像内容.算法很好地解决了水印鲁棒性和不可见性之间的矛盾.实验结果表明,该算法能够抵抗JPEG压缩、加噪和任意裁剪等攻击,具有很强的鲁棒性.A blind digital watermarking algorithm based on image content and wavelet packet transform is advanced. Firstly, the image is divided into three different areas：smooth area, edge area and texture area. Then due to the sparse representation of wavelet packet transform for image features, positions of important vision in- formation where watermarks are embedded can be founded out. The embedding strength is adaptively determined by the original image content. The proposed algorithm solved the conflict of watermark robustness and perceptual invisibility. Experimental results indicate that the algorithm is robust against attacks such as JPEG, noise and random crop.

关 键 词：数字水印 小波包变换 自适应 图像属性 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]