计算机应用

• 图形图像处理 • 上一篇    下一篇

高精密点云数据三维人头模型的变形

龚蕾   

  1. 河南科技大学电子信息工程学院
  • 收稿日期:2005-12-02 修回日期:1900-01-01 发布日期:2006-06-01 出版日期:2006-06-01
  • 通讯作者: 龚蕾

Distortion of 3D head model with high-precision point cloud data

Lei Gong   

  1. 河南科技大学电子信息工程学院
  • Received:2005-12-02 Revised:1900-01-01 Online:2006-06-01 Published:2006-06-01
  • Contact: Lei Gong

摘要: 提出了一种基于高斯变换的计算机三维头像变形算法,可以全自动产生三维头像变形效果,该算法摒弃了其他人脸变形算法中所必需的标定人脸关键点的繁琐工作,提出一种只需要设置一个中心点以及变形参数就可以将三维头像进行变形的方法, 实验结果表明,该算法所需用户控制少,方便用户对渐变精细度的控制,运算速度快,产生理想的平滑渐变,能够很好地达到预期变形效果。该工作有助于将人脸变形技术运用于三维动画,人脸识别等领域。

关键词: 物体编码, 点云数据, 高斯变换, 透视投影

Abstract: A novel approach for facial transformation based on guass transform was presented, which provided automatic metamorphosis among face images. It can eliminate Human intervention for feature specification caused by the facial marking. In addition, this method makes it possible to transform the 3D figures based on only one center and a few other parameters. The result of experiments indicates that this algorithm needs less control of users. It is convenient for users to control the definition of the gradual change. This algorithm can get a high operating speed, ideal smooth gradual change, and anticipative effect of distortion. The algorithm is excellently adaptable to the area of cartoon and face distinguishes.

Key words: object coding, point cloud data, gauss transform, perspective projection