基于分块的DWT和PCA图像水印算法  

Block-Based Image Watermarking Method With DWT and PCA

作　　者：夏维伟 陈家琪[1] 

机构地区：[1]上海理工大学光电信息与计算机工程学院,上海200093

出　　处：《电子科技》2017年第11期96-99,共4页Electronic Science and Technology

摘　　要：针对传统DWT水印算法抵抗几何攻击鲁棒性差,文中提出一种基于分块的离散小波变换(DWT)和主成分分析(PCA)相结合的水印算法。通过对载体图像进行三阶小波分解,在第3层分解的低频逼近子图上用主成分分析提取出主成分系数,将水印嵌入到提取的主成分系数中。实验结果表明,相对于传统DWT算法该算法在抵抗各类攻击均表现出良好的鲁棒性。A new watermarking algorithm based on discrete wavelet transform（DWT） and principle component analysis（PCA） is proposed to overcome the weakly robustness of traditional watermarking algorithm against geometric attacks. The third-order wavelet decomposition is applied to the original image,which principal component coefficients are extracted by the principal component analysis on the low-frequency approximation subgraph of the third layer decomposition. The watermark is embedded in principal component coefficients. The results show that the proposed algorithm has better robust against various attacks compared with traditional DWT algorithms.

关 键 词：图像水印 分块 离散小波DWT 主成分分析PCA 鲁棒性 

分 类 号：TP309.7[自动化与计算机技术—计算机系统结构]