计算机工程 ›› 2013, Vol. 39 ›› Issue (3): 250-253.doi: 10.3969/j.issn.1000-3428.2013.03.049

• 图形图像处理 • 上一篇    下一篇

基于改进ASM的三维人脸自动对齐算法

胡阳明?,周大可,鹿 乐,杨 欣   

  1. (南京航空航天大学自动化学院,南京 210016)
  • 收稿日期:2012-04-05 出版日期:2013-03-15 发布日期:2013-03-13
  • 作者简介:胡阳明(1988-),男,硕士研究生,主研方向:模式识别,图像处理;周大可,副教授、博士;鹿 乐,硕士研究生;杨 欣,副教授、博士
  • 基金项目:
    国家自然科学基金资助项目(61172135);南京航空航天大学研究生创新基地(实验室)开放基金资助项目(kfjj120211);中央高校基本科研业务费专项基金资助项目

Automatic 3D Faces Correspondence Algorithm Based on Improved ASM

HU Yang-ming, ZHOU Da-ke, LU Le, YANG Xin   

  1. (College of Automation Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, China)
  • Received:2012-04-05 Online:2013-03-15 Published:2013-03-13

摘要: 针对传统三维人脸对齐算法需要大量人工操作和数据冗余的不足,提出一种基于改进主动形状模型的三维人脸自动对齐算法。该算法对二维人脸进行快速自动特征定位,根据三维人脸柱面展开的二维纹理图进行自动特征定位并分片。采用基于平面模板的重采样算法,自动实现三维人脸顶点稠密对齐。实验结果表明,该算法在提高运算速率的同时,得到的归一化结果与原始数据仍能保持99%形状相似度。

关键词: 主动形状模型, 三维人脸, 稠密对齐, 重采样, 特征定位

Abstract: The traditional algorithm is inefficient in 3D faces point-to-point correspondence. To address this problem, an improved Active Shape Model(ASM)-based automatic 3D faces correspondence algorithm is proposed. The algorithm rapidly locates a 2D face. And accurately position the feature points in the texture maps of 3D models, which are generated by unwrapping 3D faces to 2D space. Feature location information is used to patch the unwrapping map. The resample algorithm based on the planer template is used to automatically achieve the pixel-wise correspondence between 3D faces. Experimental results show that the algorithm not only can improve the operation rate, but also the normalized results obtained remains 99% shape similarity with the original data.

Key words: Active Shape Model(ASM), 3D face, dense correspondence, re-sampling, feature location

中图分类号: