Iris Outer Boundary Localization Based on Leading Edge Technique

Abstract

In recent years, the iris biometric occupies a wide interesting when talking about
biometric based systems, because it is one of the most accurate biometrics to prove
users identities, thus it is providing high security for concerned systems. This
research article is showing up an efficient method to detect the outer boundary of
the iris, using a new form of leading edge detection technique. This technique is
very useful to isolate two regions that have convergent intensity levels in gray scale
images, which represents the main issue of iris isolation, because it is difficult to
find the border that can separate between the lighter gray background (sclera) and
light gray foreground (iris texture). The proposed method tried to find iris radius by
seeking in the two iris halves (right and left) circularly, in term of certain angles
interval for each half, to avoid the existence of the upper and lower eyelids and
eyelashes. After the two radiuses (i.e. for each half) had been determined, the iris
final iris radius would be evaluated to the minimum value of them. This method
tested on all samples of CASIAv4-Interval dataset, which consist of 2639 samples,
captured from 249 individuals, and distributed on 395 classes, the accuracy of the
testing was 100% for outer boundary detection.