A Vision-Based Hybrid Method for Eye Detection and Tracking

As one of the most salient features of human face, eyes play an important role in interpreting and understanding a person’s desires, needs, and emotional states. In this paper, a robust real-time “coarse to fine” method based on combining the appearance-based method and SVM approach was proposed for eye detection firstly, which combining the respective strengths of different complimentary techniques and overcoming their shortcomings. And then, a complementary tracker was designed for tracking eyes, which the eye model is continuously updated by having the eye successfully detected from the last Kalman tracker, to avoid error propagation with the CamShift tracker. Experimental results show that these enhancements have led to a significant improvement in eye tracking robustness and accuracy over existing eye trackers, especially under various conditions identified.