Methodology for iris segmentation and recognition using multi-resolution transform

Abstract

Iris segmentation is used to locate the valid part of the iris for iris biometrics which is an essential module in iris recognition because it defines the effective image region used for subsequent processing such as feature extraction and iris identification. A novel algorithm for efficient and accurate iris segmentation is carried out in this system. The pupil boundary is detected by applying the equation of circle by finding three points on its circumference. The reflection within the pupil region (if any) is filled by reducing the radius of the pupil one by one until it reaches to zero. Then calculating the edge points of iris boundaries (left, right, upper and lower) point by taking the fixed value from pupil circumference. The novelty here for eyelids localization can be performed by using ‘3 points marking’ for upper lid and ‘edge detector’ for lower lid. After that, eyelash removal can be done by Order — Statistic Filtering. Finally, the accurate iris edge region is fitted by calculating the point of intersection between eyelids and eye localization. After edge fitting, the curvelet transform is applied for feature extraction. The Manhattan and Euclidean Distance measures are used to measure the similarity between two images to find the best match. Here, the challenging benchmark database MMU is used for identification and verification.