基于小波系数块能量和HVS的FCM水印算法  

A FCM Watermarking Algorithm Based on Energy of Wavelet Coefficients Block and HVS

作　　者：张民[1] 郭玉彬[2] 张德伟[1] 

机构地区：[1]聊城大学计算机学院,山东聊城252059 [2]华南农业大学信息学院,广东广州510640

出　　处：《微电子学与计算机》2010年第5期139-142,共4页Microelectronics & Computer

基　　金：山东省科技攻关计划项目(GG3WZ04015)

摘　　要：为增强水印的鲁棒性,提出一种基于小波变换的改进数字水印算法.该算法将原始图像小波变换后的细节子图分割成互不重叠的系数块.通过对各系数块能量特征分析,并结合人类视觉系统特征(HVS)对特征值进行模糊c-均值聚类(FCM)分析取得更适合嵌入水印的块,在适合嵌入块上应用隶属度值进行自适应的水印嵌入.实验表明该算法有更好的鲁棒性和不可见性.To enhance the robusrness of watermark,an improved watermarking algorithm based on wavelet transform was proposed. The level detail sub-image decomposed the original image using three levels discrete wavelet transform （DWT） were divided into non-overlap blocks. The blocks which were more adaptable to embed watermark,were selected by FCM on the block energy and HVS properties,and the cluster were used as the strength of watermarking on the adaptable blocks. The experimental results show that the embedded watermark is invisible,and also very robust.

关 键 词：数字水印 人类视觉系统特征 块能量 小波变换 模糊C-均值聚类 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]