期刊文献+

3D人脸姿态校正的几何代数不变量方法 被引量:2

3D Face Pose Correction Based on Geometrical Invariance by Geometrical Algebra Theory
下载PDF
导出
摘要 多模态3D人脸曲面存在姿态自由,表情丰富的特征,这给人脸的特征提取及提高分类器的识别率带来困难,因此需要对人脸样本进行姿态矫正.本文提出的基于几何代数(Geometrical Algebra)不变量方法的3D人脸姿态校正策略,计算非标准姿态样本与参考标准样本的几何代数不变量,并且快速计算彼此不变量之间的空间几何关系,实现不同人脸曲面的空间姿态校正.实验结果表明:该方法对与标准姿态人脸样本相同、相异身份人的样本都能较好的实现姿态校正,并且几何意义直观,计算简单,校正精度高. Multi-modal 3D face surface has the characteristics of free space attitude and rich expression, which brings inconvenience and difficulties to feature extraction of face data and improve the recognition rate of classifier,it needs pose correction for face sam- ples. 3D face posture correction strategy based on the method of Geometrical Algebra invariant proposed by this paper, which calcu- lates the corresponding geometrical algebra invariant of samples for calibration and the standard samples, and estimates the spatial geo- metric relationships of invariant between each other quickly, realizes space attitude correction for different facial surface. The experi- mental results show that the method can achieve a better posture correction for the samples of people with different identity but has the standard pose face, have the features of intuitive geometric meaning,simple calculation and high correction precision.
出处 《小型微型计算机系统》 CSCD 北大核心 2015年第1期177-181,共5页 Journal of Chinese Computer Systems
基金 国家自然科学基金项目(60871093)资助 浙江省重中之重学科信息处理与自动化技术开放基金项目(20120816)资助 宁波市自然科学基金项目(2012A610050)资助
关键词 3D人脸 姿态校正 几何代数 几何不变量 3D face pose correction geometrical algebra geometrical invariance
  • 相关文献

参考文献2

二级参考文献16

  • 1梁国远,查红彬,刘宏.基于三维模型和仿射对应原理的人脸姿态估计方法[J].计算机学报,2005,28(5):792-800. 被引量:25
  • 2方惠兰,王国瑾.三角网格曲面上离散曲率估算方法的比较与分析[J].计算机辅助设计与图形学学报,2005,17(11):2500-2507. 被引量:38
  • 3刘晓宁,周明全,耿国华.基于单张二维照片的三维姿态计算[J].计算机工程,2006,32(6):232-233. 被引量:4
  • 4Chen Q, Wu H, Shimada T. A robust algorithm for 3D head pose esti- mation [ C]//Proceedings of IEEE International Conference on Multi- media Computing and Systems. Nara, 1999:697 - 702.
  • 5Bruske J, Abraham Mumm E, Pauli J, et al. Head pose estimation from facial images with subspace neural networks [ C ]//Proceedings of Internal Neural Networks and Brain Conference. Beijng, China, 1998 : 528 - 531.
  • 6Alter T D. 3-D pose from 3 points using weak-Perspective [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1994, 16 (8) :802-808.
  • 7Yao P. Evans G,Calway A. Using affine correspondence to estimate 3D facial pose [ C ]//Proceedings of the IEEE International Conference on Image Proceeding. Thessaloniki, The Hellenic Republic, 2001,3 : 919 - 922.
  • 8Mazumda Debasis R, Dutta Santanu, Mitra Soma. Automatic feaiure detection of a face and recovery of its pose [ C ]//Communicated to Journal of IETE. Washington, USA. 2003:505 - 511.
  • 9Wan Kwork Wai , Lain Kin Man , Chang Kit. An accurate active shape model for facial feature extraction [ J]. Pattern Recognition Let- ters, 2005,26(12) :2409 -2423.
  • 10Yun Hai Xiao , Hui Juan Sun , Zhi Guo Wang. Aglobalty convergent BFGS method with nonmonotone line search for non-convex minimiza- tion[ J]. Journal of Computational and Applied Mathematics, 2009, 230:95 - 106.

共引文献6

同被引文献16

引证文献2

二级引证文献5

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部