A Robust Iris Segmentation Algorithm Based on Pupil Region for Visible Wavelength Environments

In the last decade, the research on iris biometric has received increasing attention. Most of this research targets the iris recognition scenarios in constrained or controlled conditions, where considering the unconstrained environments is still needs more research work. In unconstrained environments, the sources of noise in eye regions are significantly more than constrained environments, leading to severe degradation in the iris region. As a result, iris segmentation step has a crucial significance and becomes a major issue in unconstrained iris recognition, since most of the traditional iris segmentation techniques fail under such challenging conditions. In this paper, a new segmentation algorithm is proposed to handle iris images acquired in visible wavelength environments. The proposed segmentation algorithm decreases the degradation and noise by starting from the most easily distinguishable region of the iris, which is the dark circular region called pupil. After that, the iris is localized accurately using a fast-circular Hough transform. Finally, the upper and lower eyelids and eyelashes are determined and removed from the iris region by applying a set of more suitable methods for unconstrained environments. The proposed algorithm is compared with several state-of-the-art segmentation algorithms using the UBIRIS database, and the results validate the effectiveness and stability of the proposed algorithm.