Eye gaze calibration based on support vector regression machine

Authors:

Abstract
For eye gaze tracking in the paper propose the method based on the support vector regression to build up mapping function between the iris position and the eye gaze position. In addition, we use the central coordinates of the iris as input features of the support vector regression algorithm. The experimental result shows the method for eye gaze tracking is superior to the traditional tracking algorithms making polynomial as regression model in the estimation accuracy.