一种基于聚类分析的三角网格强壮水印方法  

Cluster Analysis Based Robust Mesh Watermarking Approach

作　　者：王彦伟[1] 郑英[2] 

机构地区：[1]华中科技大学国家CAD支撑软件工程技术研究中心,武汉430074 [2]华中科技大学控制科学与工程系,武汉430074

出　　处：《计算机科学》2009年第2期251-256,共6页Computer Science

基　　金：国家自然科学基金(60604030);国家"863"基金项目(2007AA04Z136)资助

摘　　要：针对计算机图形学中常用的三角网格模型,以聚类分析方法为基础建立了一种强壮的水印方案。首次将聚类分析常用的K-Means方法用于网格水印研究,并在完成球面三角化后通过调整表面片法矢嵌入水印;在分析方法特点基础上建立了高效的数值迭代过程,提高了水印方法的效率;通过实验结果验证了方法的有效性和鲁棒性。We presented a cluster analysis based robust watermarking approach＇which can be used for watermark embedding and extraction of triangle mesh, a frequently research object in computer graphics. For the first time, we applied K-Means cluster method, which is a common method in cluster analysis, into mesh watermarking research. Watermark information was embedded by spherical triangulation of cluster result and subtle modification of facet normals. By analyzing characters of the whole process, the high efficient numerical iteration process was constructed. Discussion of the parameters and their effect used in our approach was given. And the effectiveness and robustness were testified too.

关 键 词：网格水印 聚类分析 攻击 网格对齐 网格重采样 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]