Pupil Detection in Infrared Images based on Active Contour Models

Morteza Sadri, Kaveh Kangarloo, and Fardad Farokhi


pupil, contour detection, snake algorithm, thresholding


Accurate pupil contour detection plays an important role in iris recognition and eye tracking systems. In this study a precise pupil contour detection based on snake algorithm have proposed. A camera with infrared illuminator is used. Our method consists of 4 stages; first, eyelashes are removed by applying a lowpass filter, secondly, images are going through a thresholding loop in order to satisfy GLD condition and obtain the appropriate threshold level, thirdly accumulative image of probability of pupil presence is created. In the final stage snake algorithm with 20 samples is applied to locate the accurate contour of pupil. Experimental results of the proposed method which applied to casia v3.0 iris images database, have demonstrated reasonable accuracy and speed of pupil detection.

Important Links:

Go Back