3D facial motion tracking by combining online appearance model and cylinder head model in particle filtering  

作　　者：YU Jun WANG ZengFu 

机构地区：[1]National Laboratory of Speech and Language Information Processing [2]Department of Automation,University of Science and Technology of China [3]Institute of Intelligent Machine,Chinese Academy of Sciences

出　　处：《Science China(Information Sciences)》2014年第2期272-278,共7页中国科学（信息科学）（英文版）

基　　金：supported by National Natural Science Foundation of China(Grant No.61303150);Anhui Province Innotiative Funds on Intelligent Speech Technology and Industrialization(Grant No.13Z02008);Fundamental Research Funds for the Central Universities(Grant No.WK2100100020);China Postdoctoral Science Foundation(Grant No.2012M521248);STP of Anhui(Grant No.11010202192)

摘　　要：A 3D facial motion tracking approach has been proposed based on the incorporation of online appearance model （OAM） and cylinder head model （CHM） in the framework of particle filtering. 1） For the construction of OAM, multi-measurements are infused to reduce the influence of lighting and person dependence. 2） The global motion acquired from CHM fitting is set as the initialization of OAM fitting, and the fitting result is set as the initialization of CHM in the next frame. 3） Motion filtering is applied with particle filter combined with local optimization and improved resampling. Objective between subjects evaluations show that our approach is fit to track facial motions.A 3D facial motion tracking approach has been proposed based on the incorporation of online appearance model （OAM） and cylinder head model （CHM） in the framework of particle filtering. 1） For the construction of OAM, multi-measurements are infused to reduce the influence of lighting and person dependence. 2） The global motion acquired from CHM fitting is set as the initialization of OAM fitting, and the fitting result is set as the initialization of CHM in the next frame. 3） Motion filtering is applied with particle filter combined with local optimization and improved resampling. Objective between subjects evaluations show that our approach is fit to track facial motions.

关 键 词：facial motion tracking online appearance model cylinder head model particle filtering informa-tion fusion 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]