Iris Recognition Using a Novel Segmentation Method and Fuzzy ARTMAP Neural Network

A biometric system provides automatic identification of an individual based on unique features or characteristics possessed by that person. Iris recognition is regarded as one of the most reliable and accurate available biometric systems. This paper proposes an efficient iris recognition system that employs a novel method to localize the iris region on the eye image based on approximating the iris radius using the pupil characteristics and cumulative sum based gray change analysis technique to extract features from the normalized iris template. Fuzzy ARTMAP neural network is also applied to classify the iris codes. The results of simulations on a set of 756 eye images illustrated a fast, accurate and noise resistant personal identification system. The proposed system achieved zero false acceptance rate using 1800-bit binary iris codes and recognized all authorized users with 100% accuracy.