Title: A cross-correlated feature fusion technique for multimodal biometric system

Authors: R. Vinothkanna; K. Santhi

Addresses: Department of EEE, Guru Nanak Institute of Technology, Hyderabad, India ' Department of ECE, Guru Nanak Institutions Technical Campus, Hyderabad, India

Abstract: Biometrics is one important class of security systems which is gaining widespread significance in recent times. Almost all real-time applications in practice today ranging from banking to public and private sectors utilise biometric system of authentication for employee identification, client services. The biometric data are usually stored in a large database which bears a direct consequence on the security of stored biometric features. The proposed system is a multimodal biometric system developed for passport verification incorporating three biometric traits such as iris, fingerprint and finger vein. A multi-resolution frequency domain approximation technique has been utilised for efficient feature extraction followed by a fusion rule for matching to decide upon the authenticity of the use. The proposed system has been implemented using a real-time fingerprint scanner and the classification rate observed from the experimental results indicates the superiority of the developed system.

Keywords: fingerprints; multimodal biometrics; multimodal features; feature extraction; pattern matching; cross-correlation; feature fusion; security; authentication; passport verification; passports; iris recognition; finger veins.

DOI: 10.1504/IJBET.2017.082660

International Journal of Biomedical Engineering and Technology, 2017 Vol.23 No.2/3/4, pp.180 - 190

Received: 12 May 2016
Accepted: 27 Sep 2016

Published online: 24 Feb 2017 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article