基于最小二乘网格的模型变形算法.pdfVIP

计算机辅助设计与圆形学学报 第 22 卷第 5 期 Vo l. 22 No.5 2010 年 5 月 Journal of Computer幡Aided Design . Computer Graphics May 2010 基于最小工乘网格的模型变形算法 王仁劳1\ 许秋儿 1) (浙江万望学院计算机与倍息学院 宁波 3剖10ω02幻7) 的(浙江大学计算机科学与技术学院 杭州 310027) (renfangwang@gmai1. com) 摘 要g 针对白白变形技术难以保持模型细节的问题，提出一种基于最小工乘网格的模型变形算法.通过顶点位置 约束的全局拉普拉斯光顺分解tf1表示模裂低频信号的最小二乘网格，并求出高频信号在该网格上的编码z 通过用户 交直，基于均值坐标对最小二乘网格进行自由变形;根据最小工乘网格各顶点处局部标架在变形时的几何变换求出 变形屑的高频编码，通过解码求出变形后的网格模型.实验结果表明，该算法简单、高效且便于用户交豆，有效地保 持了模型的几何细节. 关键诩g 最小二乘网格p 高频信号p 均值然标g 自由变形p 局部标架 中阁法分提高警: TP391 Shape Deformatioo ßased 00 Least-Square Mesh ll l 2J Wang Renfang • Xu Qiue户， Wang Qin ) , Liu Yunpeng!) , and Zhang Sanyuan 1) (College 01 Com户uter Scíence and Inlormation Technology , Zhejiang Wanli Uníversíty. Níngbo 315100) Z) (College 01 Computer Science and Technology , Zhejiang Universíty , Hangzhou 310027) Abstract: The main drawback of freeωform deformation method is not able to preserve the features of shape. To address it , an efficient shape deformation algorithm is proposed based on léast-叫uare mesh. By using the global Laplacian smoothíng with positional constraints. the least square mesh (LSM) including only low唰frequency component of a shape is first extracted , and the high-fr叫uency signals of the shape are encoded based on LSM. According to user editing and mean value coordinates , LSM is then deformed with fre e-form deformation method. By geometric transformatíons at each vertex of LSM , the codes of the deformed high栅frequency components are determíned , and by decoding them , the deformed shape is finally achieved. Some experimental results demonstrate the algorithm is very fast , easy to manipu