Robust Gaze Estimation via Normalized Iris Center-Eye Corner Vector

Gaze estimation plays an important role in many practical scenarios such as human robot interaction. Although high accurate gaze estimation can be obtained in constrained settings with additional IR sources or depth sensors, single web-cam based gaze estimation still remains challenging. The purpose of this study is to improve the robustness of pupil center-eye corner(PC-EC) vector based gaze estimation method against slightly head movement with a single, low cost web-cam. Firstly, reliable facial features and pupil centers are extracted. Then, a normalised iris center-eye corner(NIC-EC) vector is proposed to improve the robust and accuracy of gaze estimation. Finally, an interpolation method is employed for the mapping between constructed vectors and points of regard(PoG). Experimental results showed that our method can achieve better performance than the PC-EC based gaze estimation method under slight head movements. The average accuracy of gaze estimaion with calibration and testing is around 1.66 and 2.68 respectively.