A novel remote eye gaze tracking approach with dynamic calibration

Point of gaze estimation is the most important part of remote eye gaze tracking techniques. Though there are various point of gaze detection and estimation methods, few of them can satisfy the expectation of widely use. One primary reason is lack of accurate calibration method. In order to deal with this problem, an adaptive calibration technique is proposed based on cross-ratio. It can compensate estimation bias much better. Also, efficient remote eye gaze tracking parameters estimation procedures are employed in this paper. The eye gaze tracking approach in this paper allows users to have free head motion and has high accuracy rate, and it is based on infrared radiation (IR) light sources reflection on cornea. Experimental results demonstrate that considerable improvement can be achieved.