Iris Recognition System for Secure Authentication Based on Texture and Shape Features

One of the most accurate biometric authentication methods is iris pattern. It has the advantages of being stable, contactless and no user's previous knowledge is required. This paper presents an iris recognition system for user authentication. To design the proposed iris authentication system we reviewed and evaluated four iris pattern recognition features including Histogram of Oriented Gradients (HOG), combined Gabor and Discrete Cosine Transform (DCT), and Grey level Co-occurrence Matrix (GLCM). The system was tested using UBIRIS.v1 IRIS dataset and the results showed that GLCM gives the largest Euclidean distance between two iris images for two different users, which is higher than using combined features. Moreover, GLCM gives the highest recognition accuracy using Logistic Model Trees (LMT) classifier. Accordingly, GLCM is regarded the most discriminative and the most effective technique for the proposed iris authentication system.