An adaptive calibration applying SLRA for gaze tracking

Video-oculography (VOG) is one of the most attractive gaze tracking techniques because of its convenience and simplicity, the purpose of the calibration of VOG gaze tracking system is to find the mapping function between the point of gaze (POG) in the equipment coordinates and the eye movement features in the image coordinates. In this paper, an adaptive calibration applying Stepwise Linear Regression Analysis (SLRA) for VOG gaze tracking systems will be introduced. We use a high order polynomial expression as a primary mapping function, then estimate the significance of the variables of the mapping function, and get the best mapping function by removing non-significant variables. The experimental result shows that the average error of this method is less than 0.5°. The proposed method has better accuracy and stability compared with the traditional quadratic polynomial mapping method.