Pupil localization by ASM and regional gray distribution

This paper presents a new approach for pupil localization. The whole system can be divided into two stages: eyes detection and pupil localization. In the first stage, ASM method is used to locate the facial feature points, in which Adaboost algorithm is used to detect human face and provide the original searching position. In the second stage, gray distribution of the eyes is analyzed. The local area with lowest gray intensity is assumed the candidate of pupil. Finally, similar local area analysis technique is employed to adjust the position of the candidate pupil area. Experimental results validate the effects of our approach.