Polymer thick film sensors for embedded Smartcard Biometrics and identity verification

Abstract

Smartcards offer robust mechanisms for electronic demonstration of identity, and as a result are an increasingly important tool of e-commerce and information security. The identity of an authorised individual and their smartcard are, however, weakly bound and the presence of a legitimate and card-holder can not be assured. Hence, there is growing interest in the use of biometrics to strengthen the association between card and card-holder, although the emphasis has predominantly been to use external components such as sensors or processing elements. Such approaches expose the risk of eavesdrop and replay attacks, and because of this it is describe to incorporate all elements of an identity-verification system on-card.

Potential discriminatory characteristics are limited to the measurable interactions between an individual and smartcard. Hence, the emerging field of biometrics is comprehensively reviewed for plausible mechanisms of demonstrating identity. Spatial finger characteristics such as fingerprints, finger-geometry and finger-crease pattern emerge as the most likely approaches.

Sensors integrated onto smartcards are required to be mechanically flexible, robust and of low-cost, and such conditions are satisfied by a class of sensors known as polymer thick film (PTF). Piezoelectric and piezoresistive PTF pressure sensors are considered, and shown to exhibit sensitivity to loads applied elsewhere on the card. This effect renders the concept of using PTF pressure sensors to capture spatial finger characteristics infeasible.

Nevertheless such sensing mechanisms can be exploited to capture temporal, rather than spatial interactions, and a novel approach to identity verification is proposed and demonstrated. This approach is based upon the pressure response of finger-taps, and an experiment involving 34 participants demonstrates significant discrimination between individuals. An investigation into suitable verification functions reveals that an equal error rate of 2.3% is achievable under controlled laboratory conditions. Necessary enrolment and verification algorithms are implemented on a typical smartcard platform, and found to execute within 3.1 and 0.12 seconds, respectively. Full compliance with the mechanical, computational, and economical constraints of a smartcard is consequently demonstrated.

More information

Identifiers

Catalogue record

Export record

ASCII CitationAtomBibTeXData Cite XMLDublin CoreDublin CoreEP3 XMLEndNoteHTML CitationHTML CitationHTML ListJSONMETSMODSMPEG-21 DIDLOpenURL ContextObjectOpenURL ContextObject in SpanRDF+N-TriplesRDF+N3RDF+XMLRIOXX2 XMLReferReference ManagerSimple Metadata