Biometric Feature Extraction for Iris Scans

Abstract

This paper generally seeks to demonstrate understanding and use of iris scans for biometric identification. Salient
features of iris are identified, their extracts analyzed critically by use of a software. The software for extraction of the
features is developed on MATLAB platform. Already prepared iris scans are used due to lack of scanners and/or very
high resolution cameras that would have otherwise assisted to obtain the required distinct features for comparison. The
program developed would compare two iris images a time and determine if they came from same individual; what we
are referring to as matching. A computer vision algorithm based on The Hough Transform is made use of in determining
circles present in the image which are essentially the iris and pupil. The region of interest from which the unique features
are obtained would be located between papillary and limbic boundaries. Papillary (between iris and pupil) while limbic
boundary is the one demarcating the iris and sclera. The comparison entails first determining codes of the two iris images
through generation of separate biometric template for them then using encoding procedure. Hamming distance calculation between the iris codes determines if they match.