计算机工程 ›› 2008, Vol. 34 ›› Issue (12): 236-238.doi: 10.3969/j.issn.1000-3428.2008.12.083

• 开发研究与设计技术 • 上一篇    下一篇

一种构造人脸原型的Face-LS算法

蒙应杰,王 维,张 文,郭喜平   

  1. (兰州大学信息科学与工程学院,兰州 730000)
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2008-06-20 发布日期:2008-06-20

Face-LS Algorithm for Constructing Face Prototype

MENG Ying-jie, WANG Wei, ZHANG Wen, GUO Xi-ping   

  1. (School of Information Science and Engineering, Lanzhou University, Lanzhou 730000)
  • Received:1900-01-01 Revised:1900-01-01 Online:2008-06-20 Published:2008-06-20

摘要: 将Loop细分原则引入语音驱动的语音动画合成系统中,提出并设计一种基于局部细分模型构造人脸原型的Face-LS算法。算法利用局部细分原则在脸部网格的局部区域上进行迭代插值,经插值细分后在降低整体网格密度的前提下,导致不同脸部区域特征点密度不同,使重要区域具有较高密度,以此降低脸部原型的褶皱度。通过仿真试验验证该算法具有较好的性能。

关键词: 语音动画, 人脸原型, 局部细分

Abstract: This paper introduces the Charles Loop’s scheme into the synthesis of realistic animated facial images drove by sound-driver, and proposes a Face-LS algorithm based on location subdivision module. This algorithm does iterative interpolating ghost points on location using location subdivision principle, in order to hold different density on different region, which can prevent distortions and decrease the density of the whole mesh. Simulation experiment shows that the algorithm has better performance.

Key words: speech animation, face prototype, location subdivision

中图分类号: