Design and Implementation of an Iris Recognition System

Abstract

Hough transform is still the most commonly used method in many Iris Recognition
Systems (IRSs) used in large companies and airports around all over the world. However,
studies implemented demonstrated that Hough transform is a time consuming method.
Moreover, the algorithm should be repeated twice (adding further time consuming); one for
the isolation of the iris region and the other for the pupil region through the detection of the
radius and the centre of both regions.
In this paper, a new technique to isolate the iris and the pupil regions for an IRS has
been proposed. This technique, by considering the centre of the iris is exactly that of the
pupil, applies two new algorithms, one to find the radius and the centre of the pupil and the
other to find the radius of the iris. These algorithms make use of the intensity values of the
pupil and the iris areas through setting different thresholds for their isolation. The new
technique is programmed using MATLAB 6.5.
The results of this study have shown that the time required to isolate the iris region
and the pupil region can be reduced rapidly using this new technique. Where it requires only
(3) seconds to isolate both the iris and the pupil regions on a Pentium IV PC (1.7 GHz) using
this new technique. This new applied technique to isolate the iris region and the pupil region
has been tested on a 100 images database, all reporting a zero failure rate.