An Elliptic Curve Algorithm for Iris Pattern Recognition

Abstract

This paper presents an iris recognition system with a self-developed elliptic curve algorithm or hash. The system was written in Python with the OpenCV library. This iris recognition system is faster than standard systems because the extraction of the iris was done on a grayscale image. To test the elliptic curve hash, iris recognition system, and computation time, images from the UBIRIS database were run through the system, and their hashes were collected. Statistical analysis shows that the elliptic curve hash has more entropy and is quicker than the standard MD5 and SHA-1 hashes at an alpha of 0:01. Furthermore, the system has an accuracy rate of 99:5%.