Pupils Tracking Using Multiple Coordinating Particle Filters

Yusuke OHTANI  Kazuhiro FUKUI 

Publication
D - Abstracts of IEICE TRANSACTIONS on Information and Systems (Japanese Edition)  Vol.J95-D  No.6  pp.1380-1389
Publication Date: 2012/06/01
Online ISSN: 1881-0225
Print ISSN: 1880-4535
Type of Manuscript: PAPER
Category: 
Keyword: 
pupil tracking,  particle filter,  separability filter, 

Full Text(in Japanese): PDF(2.8MB)

Summary: 
In this paper, we propose multiple coordinating particles filters to achieve a stable pupil tracking against the changes of head pose and gaze directions. Particle filter is an effective algorithm for tracking targets, such as pupils, with complex movements. However, a straightforward approach using the original particle filters lacks the stability of tracking, since it track two pupils independently. To improve the stability, we focus the coordinated motion of two pupils. In our filters, the next state of a pupil is estimated by not only itself state but also that of another pupil. We demonstrate the effectiveness of the proposed filters through evaluation experiments.