A Novel Behavioural Biometric Technique for Robust User Authentication. Issue 5 (3rd September 2017)

• Record Type: Journal Article
• Authors: Bhardwaj, Ishan; Londhe, Narendra D.; Kopparapu, Sunil K.
• Is Part Of: IETE technical review. Volume 34:Issue 5(2017)
• Page Start: 478
• Page End: 490
• Publication Date: 2017-09-03
• Subjects: Authentication -- Biometrics -- Classification -- Feature subset selection -- Fingerprint dynamics -- Pattern recognition; Telecommunication -- Periodicals; Electronics -- Periodicals; Electronics; Telecommunication; Periodicals; 621.38
• DOI: 10.1080/02564602.2016.1203271
• Languages: English
• ISSNs: 0256-4602

Abstract

ABSTRACT: In this paper, we propose a novel approach termed as f ingerprint d ynamics, targeted for multi-instance authentication system. Inspired by the concept of keystroke dynamics, the technique analyses the behavioural characteristics associated with multi-instance fingerprint acquisition. To capture the user data, an acquisition set-up (hardware) is devised. Data-set of genuine users and imposters are created using time-derived information, collected for unique fixed length sequences by 32 volunteers. Systematic evaluation of features and feature subset selection is performed. Goal of this paper is to determine potential of fingerprint dynamics and establish it as a successful biometric trait for the purpose of robust user authentication. The technique can be implemented as a part of multifactor authentication system or standalone alternative. Statistical analysis of classifier outputs for different kernels, by performing classification cross validation, and assessment of verification accuracies confirms the discriminating capability of proposed technique.