1051-8215 (c) 2015 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information. This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TCSVT.2016.2527299, IEEE Transactions on Circuits and Systems for Video Technology 1 A person authentication system based on RFID tags and a cascade of face recognition algorithms Filippo Battaglia, Giancarlo Iannizzotto, Lucia Lo Bello Abstract—Face recognition represents an appealing solution for biometrics-based unobtrusive and flexible person authentica- tion. However, most solutions proposed in the literature suffer from some significant drawbacks, such as, high computational complexity, the need for a centralized biometrics database (which is not desirable, due to widespread international provisions discouraging collections of sensitive personal data) and limited scalability on a large number of enrolled subjects. We propose a novel person authentication solution based on a cascade of face recognition and pattern matching algorithms, that not only provides for high reliability and robustness against impostors, but also stores in a personal RFID tag all the needed individual bio- metrics information of the user, who therefore always remains in control, and has the exclusive availability, of such sensitive data. The paper describes the proposed approach, called RFaceID, and discusses its performance in terms of the ratio between false acceptance rate and false rejection rate and in terms of authentication time, when applied to the VidTIMIT, Extended Yale B and MOBIO widely adopted face databases. Index Terms—Biometrics, face recognition, RFID tags EDICS Category: Face Recognition, Biometrics, Authentication I. I NTRODUCTION The interest towards Biometric authentication for high se- curity protection systems has been consistently increasing over the last decade. Several realistic applications have been introduced, under the strict constraint that the selected physi- ological or behavioral characteristics of the subject cannot be stolen or imitated [1]. Face characteristics are among the most frequently adopted features, as they can be easily “sensed” from a distance of a few meters without requiring direct contact or close-up image acquisition [2]. Moreover, local as well as transnational regulations often explicitly enforce that people expose their face while approaching guarded gates or otherwise accessing controlled areas. However, face characteristics are considered as sensitive personal data in several countries and therefore their collection is strongly discouraged or even prohibited. Several provisions enacted by the responsible authorities (e.g., art. 29 Working Party of the European Commission for the protection of biometric data), advise against the use of a centralized server in these kind of Manuscript received June 30, 2015; revised November 2, 2015; accepted January 8, 2016. Giancarlo Iannizzotto is with Department of Cognitive Sciences, Education and Cultural Studies, University of Messina, Italy, e-mail: {ianni@unime.it} Filippo Battaglia and Lucia Lo Bello are with DIEEI, Department of Electrical, Electronic and Computer Engineering, University of Catania, Italy, e-mail: {filippo.battaglia@dieei.unict.it}, {lucia.lobello@unict.it} systems, except for few specific purposes. The EC authority considers advisable that “biometric systems are based on the reading of biometric data stored as encrypted templates on media that are held exclusively by the relevant data subjects (e.g. smart cards or similar devices)” [3]. Therefore the approaches storing the biometric data in the authentication system (as most works in the literature do) should be avoided. In order to overcome the issues described above, in this work we introduce RFaceID, a novel two-factor authentication architecture based on an RFID tag and a cascade of two face recognition stages. The first stage uses the Two-Dimensional Principal Component Analysis (2DPCA) [4], while the second is based on the Speeded-up Robust Features Detector (SURF) [5] pattern matching algorithm. The main value-added is threefold. First, a technique based on the novel BestPoint model to jointly calculate the optimal parameters for the two stages, which allows a very high recognition rate together with an extremely low false acceptance rate. Second, the proposed approach does not require a centralized database storing the biometric data of all the authorized subjects. It only relies on the personal biometric information stored into the RFID tag, which always remains available exclusively to the user. Third, RFaceID is devised to work on images captured at a very low resolution, compatible with the storing capacity of the small memories (8-32 Kbytes) of the passive RFID tags currently on the market, with improved accuracy over the existing state of the art (both in terms of false acceptance rate and of false rejection rate) even in presence of largely varying illumination. This paper is organized as follows. Section II provides a description of related works and addresses their limitations. Section III proposes a relation between face recognition and face authentication algorithms, thus paving the way for the proposed authentication approach, which is introduced in Section IV. Sections V and VI describe the enrollment and the authentication phases, respectively. Section VII presents a testing protocol developed in order to assess the performance of the proposed approach and compares its results with those produced by a recent state-of-the-art algorithm based on Gabor Disparity [6]. Section VIII demonstrates the superiority of RFaceID with respect to the VisilabFaceRec algorithm, which shares some structural similarities with RFaceID and was recently presented in [7]. Finally, Section IX concludes the paper and gives hints for future work. II. RELATED WORK A significant effort has been made over the years to develop several template matching algorithms for face recognition Copyright c 2016 IEEE. Personal use of this material is permitted. However, permission to use this material for any other purposes must be obtained from the IEEE by sending an email to pubs-permissions@ieee.org.