A novel pupil detection algorithm for infrared eye image

Abstract

In order to improve the accuracy of the eye gaze tracking technology, this paper presents a novel pupil detection algorithm based on Hough Transform with edge gradient direction information. According to the pupil pixel features in the context of infrared corneal reflection, taking advantage of two-dimensional Hough Transform, combining the edge gradient direction and a fixed range of discrete pupil radius, this algorithm counts parameter space discrete transform points to locate pupil center. It effectively filters out noise, reduces the discrete transform point statistics and calculates pupil's parameters. Experimental results show that the algorithm has a higher accuracy and real-time than the previous models.