This paper represents a robust Facial feature tracker based on Particle filters which is applied on sequential images, have been passed through symmetric filter. The aim of SF(symmetric filter) is to highlight the selected features of face. Our method is applied on Cohen-Kanade database which contains 6 basic emotions of different faces. Also distributions and creating template vectors of features are not in the same way for all features. Our algorithm is robust in presence of illumination changes. While occurring temporary occlusion, the aim is to recover the feature points not to miss them. Recovering points of features is based on estimating them according to coordinates of reference points which have not been occluded.
5th Mosharaka International Conference on Communications, Propagation, and Electronics (MIC-CPE 2012)
Congress
2012 Global Congress on Communications, Propagation, and Electronics (GC-CPE 2012), 3-5 February 2012, Istanbul, Turkey
Pages
21-25
Topics
Monte Carlo Simulation and Forecasting Model Development
ISSN
2227-331X
DOI
BibTeX
@inproceedings{276CPE2012,
title={Particle filter based tracking of facial features enhanced by symmetrical mask in presence of occlusion},
author={Mohamadreza Nikfard, and Hadi Seyedarabi, and Ali Aghagolzadeh},
booktitle={2012 Global Congress on Communications, Propagation, and Electronics (GC-CPE 2012)},
year={2012},
pages={21-25},
doi={}},
organization={Mosharaka for Research and Studies}
}
