首页 | 本学科首页   官方微博 | 高级检索  
     

基于多点模型和改进BFGS的人脸姿态估计
引用本文:邱丽梅,李青虹. 基于多点模型和改进BFGS的人脸姿态估计[J]. 计算机应用与软件, 2011, 0(9)
作者姓名:邱丽梅  李青虹
作者单位:三明学院物理与机电工程系;
基金项目:福建省教育厅科技项目(JK2009037)
摘    要:针对人脸姿态估计往往存在姿态解不稳定和不能唯一确定人脸三维空间姿态的缺陷,准确提取人脸特征点及进行相应特征点深度值估计后,以人脸的多个特征点建立人脸模型,并利用人脸特征点近似估计人脸姿态,通过改进的BFGS(Broyden-Fletch-er-Goldfarb-Shanno)算法精确估计三维人脸空间姿态。实验结果表明,该方法不仅可以获得稳定和唯一的3D人脸空间姿态,而且与同类方法比较具有良好的姿态估计精度。

关 键 词:人脸姿态估计  多点模型  特征点  BFGS  

FACE ATTITUDE ESTIMATION BASED ON MULTI-POINTS MODEL AND MODIFIED BFGS
Qiu Limei Li Qinghong. FACE ATTITUDE ESTIMATION BASED ON MULTI-POINTS MODEL AND MODIFIED BFGS[J]. Computer Applications and Software, 2011, 0(9)
Authors:Qiu Limei Li Qinghong
Affiliation:Qiu Limei Li Qinghong(Department of Physics,Mechanical and Electric Engineering,Sanming University,Sanming 365000,Fujian,China)
Abstract:To overcome the defects in face attitude estimation that there sometimes has the unstable attitude solution and that the 3D space face attitude can not be uniquely determined,by using the modified algorithm of BFGS the 3D space face attitude was able to be estimated accurately after the face feature points were extracted accurately and were estimated the depth value of their corresponding feature points,establishing the face model with multiple face feature points as well as estimating the face attitude app...
Keywords:Estimation of face attitude Multi-points model Feature points BFGS  
本文献已被 CNKI 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号