CONCURRENCY AND COMPUTATION: PRACTICE AND EXPERIENCE Concurrency Computat.: Pract. Exper. (2011) Published online in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/cpe.1797 SPECIAL ISSUE PAPER Mobile device access control: an improved correlation based face authentication scheme and its Java ME application Kai Xi 1 , Jiankun Hu 1, * ,† and Fengling Han 2 1 School of Engineering and Information Technology, UNSW@ADFA, Canberra, Australia 2 School of Computer Science and IT, RMIT University, Melbourne, Australia SUMMARY This paper investigates face authentication based access control solutions for camera-equipped mobile devices. A new hierarchical correlation based face authentication (HCFA) scheme is proposed, which suits resource-constrained mobile devices such as mobile phones and personal digital assistants. The idea of HCFA is conducting a partial correlation output peak analysis (analyze the relationship between each cross- correlation output peak generated from selected sub-regions of a face), in conjunction with conventional direct cross-correlation methods. The experimental results on the public domain database demonstrate that the proposed scheme achieved better performance than that of the conventional direct correlation based schemes. Furthermore, HCFA was implemented on the Nokia S60 CLDC emulator using Java ME (previ- ously J2ME) programming technology in order to test the applicability and implementability. The test results show that the proposed algorithm is implementable on mobile devices. It not only shortens processing time but also reduces resource demand significantly, compared with the direct correlation algorithms. Copyright © 2011 John Wiley & Sons, Ltd. Received 29 November 2010; Revised 13 May 2011; Accepted 15 May 2011 KEY WORDS: face authentication; access control; mobile biometric; correlation pattern recognition; Java 1. INTRODUCTION Nowadays, mobile devices have been widely used all around the world. With advances of embed- ded technologies and strong driving force from the consumer market, the technologies for mobile devices have moved towards an era of convergence. As a result, a mobile phone is no longer a simple electronic device used solely for voice communication. Modern mobile phone may possess much additional functionality, such as camera, radio, MP3 player, web browser, gaming terminal, GPS navigator and TV. As the necessity and importance of mobile devices are increasing rapidly, there is a trend that mobile devices carry more sensitive information where such information can be contacts’ details, schedules, minutes, trading secrets or even classified documents. It is consumer’s desire that mobile devices not only support a growing number of applications but also secure to use. When mobile devices are lost or stolen, stored sensitive information may fall into the wrong hands. Such security problems have posed a grave threat to the current mobile environment. It urges stronger protection mechanisms against information leakage and illegitimate use of mobile devices. Biometric techniques offer a natural and reliable option for recognizing individuals. Biometric- based security systems identify a person by his/her physiological characteristics (like fingerprint, *Correspondence to: Jiankun Hu, School of Engineering and Information Technology, UNSW@ADFA, Canberra, Australia. † E-mail: J.Hu@adfa.edu.au Copyright © 2011 John Wiley & Sons, Ltd.