Improved greedy snake model for detecting accurate pupil contour

Accurate pupil contour detection is vital for iris-based biometric systems. In this paper, a new algorithm for detecting precise pupil contour is proposed. An improved greedy snake model is presented, wherein new external energy function is designed. This external energy is extracted based on the angular integral projection function (AIPF), which is a general function to perform image integral projection along angular directions. First, the approximate pupil center is obtained as the center of mass of the binarized iris image, then the pupil circular contour is localized. Finally, the accurate pupil contour is detected by employing the improved greedy snake model being initiated on the detected pupil circle. Experimental results on CASIA V3.0 iris image database, show that the improved greedy snake model is comparable with conventional greedy snake model, and the algorithm outperforms other well-known existing methods for iris segmentation in both terms of accuracy and processing time.