Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
4970310 | Pattern Recognition Letters | 2016 | 13 Pages |
Abstract
In this paper, we present the evolution of the open source iris recognition system OSIRIS through its more relevant versions: OSIRISV2, OSIRISV4, and OSIRISV4.1. We developed OSIRIS in the framework of BioSecure Association as an open source software aiming at providing a reference for the scientific community. The software is mainly composed of four key modules, namely segmentation, normalization, feature extraction and template matching, which are described in detail for each version. A novel approach for iris normalization, based on a non geometric parameterization of contours is proposed in the latest version: OSIRISV4.1 and is detailed in particular here. Improvements in performance through the different versions of OSIRIS are reported on two public databases commonly used, ICE2005 and CASIA-IrisV4-Thousand. We note the high verification rates obtained by the last version. For this reason, OSIRISV4.1 can be proposed as a baseline system for comparison to other algorithms, this way supplying a helpful research tool for the iris recognition community.
Keywords
Related Topics
Physical Sciences and Engineering
Computer Science
Computer Vision and Pattern Recognition
Authors
Nadia Othman, Bernadette Dorizzi, Sonia Garcia-Salicetti,