Biometric Authentication using Iris Recognition
Duration: March-April 2015
Status: Completed
Members: Pranav Sodhani, Akshat Bordia
Guide: Dr. M.K. Bhuyan
Overview
The objective of the project entitled, "Biometric Authentication using Iris Recognition" is to develop an authentication system using iris as a biometric. The proposed model dubbed eye-RIS (short for eye Recognition of Iris System) takes input a close-up view of human eye and employs Daugman's Integro-Diff.erential Operator for segmenting iris and pupil. Once the iris region is segmented, it is normalized into a rectangular block with fi.xed dimensions to account for imaging inconsistencies. The feature extraction process is carried out using a 3-level wavelet decomposition strategy and binarizing the obtained vector on the basis of its sign (1, if positive; else 0). For authentication, hamming distance between the feature vectors of the considered templates is computed and thresholded for decision making purpose. The UBIRIS database has been used for training and testing. We have also developed a MATLAB based GUI for user verifi.cation using iris as a password.