Digital holographic-based cancellable biometric for personal authentication. (24th March 2016)

• Record Type: Journal Article
• Authors: Verma, Gaurav; Sinha, Aloka
• Is Part Of: Journal of optics. Volume 18:Number 5(2016:May)
• Publication Date: 2016-03-24
• Subjects: digital holographic technique -- cancellable biometric -- fingerprint -- correlation; Optics -- Periodicals; 535.05
• DOI: 10.1088/2040-8978/18/5/055705
• Languages: English
• ISSNs: 2040-8978

Abstract

Abstract: In this paper, we propose a new digital holographic-based cancellable biometric scheme for personal authentication and verification. The realization of cancellable biometric is presented by using an optoelectronic experimental approach, in which an optically recorded hologram of the fingerprint of a person is numerically reconstructed. Each reconstructed feature has its own perspective, which is utilized to generate user-specific fingerprint features by using a feature-extraction process. New representations of the user-specific fingerprint features can be obtained from the same hologram, by changing the reconstruction distance ( d ) by an amount Δ d between the recording plane and the reconstruction plane. This parameter is the key to make the cancellable user-specific fingerprint features using a digital holographic technique, which allows us to choose different reconstruction distances when reissuing the user-specific fingerprint features in the event of compromise. We have shown theoretically that each user-specific fingerprint feature has a unique identity with a high discrimination ability, and the chances of a match between them are minimal. In this aspect, a recognition system has also been demonstrated using the fingerprint biometric of the enrolled person at a particular reconstruction distance. For the performance evaluation of a fingerprint recognition system—the false acceptance ratio, the false rejection ratio and the equal error rate are calculated using correlation. The obtained results show good discrimination ability between the genuine and the impostor populations with the highest recognition rate of 98.23%.