An efficient iris localization algorithm based on standard deviations

There has been a rapid increase in the need of accurate and reliable personal identification technologies in recent years. Among all the biometric techniques known, iris recognition is taken as one of the most promising methods, due to its low error rates without being invasive. Usually an iris recognition system consists of four steps: image acquisition, preprocessing, feature extraction and identification or verification. Among these steps, iris localization is a necessary and important step in iris preprocessing. In order to be more feasible in real world application environment, the performance is a key factor. In this paper, we propose an efficient localization algorithm using standard deviation which is optimized for performance. Overall it achieves a promising result on various iris datasets compared to previous work. Besides, our method gets 52% execution time deduction compared to a traditional implementation reference for the localization.