Precise and Robust Iris Segmentation in Specular Noisy Iris Images for Iris Biometrics

Abstract

Precise and Accurate segmentation of iris itself is the major challenge encountered in the process of successful implementation of iris recognition system. Number of researchers invented numerous methods to segregate the iris portion from the captured image but noises like occlusions, eyelids and specular reflections are throwing challenges in the extraction of iris part from the eye image. Most of the algorithms used in segmentation of iris are based on edge information. Here in this proposal, we are bidding an automated iterative active contour model on the database provided by CASIA- Iris-1000, for the segmentation of the iris portion. Proposed segmentation algorithm resulted 99.6% accuracy in segregating the iris on the challenging CASIA_Iris_1000.