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

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.