A Noval Saliency Based Iris Segmentation Framework

Abstract

Iris recognition is regarded as the most reliable and accurate biometric identification system available. In Iris recognition system, iris segmentation is the most time consuming and critical step, in concern to the accuracy of the process. This is a new method of iris segmentation based on the saliency map of an image. In this paper, first the saliency map which provides the visually important region has been designed to specifically locate the iris region. Then a color based masking is done to get the iris boundary. Since this method is computationally simple, a fast and reliable method which segment iris has been proposed. The reliability of the method has been checked with different eye images. The result also show that the proposed model is faster compared to the existing methods and gives a comparable accuracy.