Abstract
The use of mobile devices is continuously growing, together with their ability to store and exchange sensitive data. This has spurred the interest, from one side, in exploiting their apparent vulnerabilities, and on the other side, in protecting users from any malfeasance. In this scenario, it has become essential to develop and deploy secure access and identification protocols on mobile platforms. Another emerging aspect that also needs attention bears on the commercial and social use of identity management systems. Towards such ends, this paper proposes using biometrics as the technology of choice, and describes FAME (Face Authentication for Mobile Encounters). FAME is an embedded application for mobile devices that provides both verification and identification, including identity management to support social activities, e.g., finding doubles in a social network. Such functionalities are implemented subject to privacy and security concerns. FAME cascade architecture is modular. The single modules perform image acquisition, anti-spoofing, face detection, face segmentation, feature extraction, and face matching. FAME also provides continuous recognition and best biometric sample selection, that both accommodate varying pose and illumination. All procedures are optimized for mobile use, using low-demanding and computation-light algorithms. Preliminary questionnaires administered to users engaged with using FAME show user acceptance, ease of use, and reliability.
