A feature-based color threshold selection approach to pupil tracking

摘要

Pupil tracking can be achieved by intrusive methods (where equipment is on or inside the user) and non-intrusive methods. Intrusive methods are not preferred because they are frequently cumbersome and require a large investment in equipment. One non-intrusive approach to eye-gaze tracking involves active infrared illumination. This research investigates a novel approach to bright-pupil tracking that takes advantage of the data resolution provided by color. This method is shown to be functional and potentially robust.