Abstract

In this paper we proposed a pupil tracking method based on particle filtering in active infrared source gaze tracking system. We constructed a pupil target model with the combination of gray characteristic and shape characteristic so that the target and the background can be distinguished easily. Pupil motion model which can predict pupil location and shape feature in the next frame was built. Gray histogram similarity and geometrical shape similarity were combined to build pupil observation model, which can improve the credibility of particles. In this paper we also showed the experiment results. The applications in interactive graphical system had verified that our method was effective and feasible.

Key words: Gaze tracking, eye characters, particle filtering, face location, pupil detection, Purkinje segment.