User-Centric Biometrics: Authentication in a Self-Service Environment

Biometric authentication is the process of establishing an individual's identity based on measurable characteristics of their anatomy, physiology or behavior. Biometrics do not share many of the limitations of traditional authentication mechanisms, as the characteristics used for biometric authentication cannot be lost, forgotten or easily replicated. Despite these advantages, there are unresolved problems with the use and acceptability of biometrics and the technology has not seen the strong uptake that many predicted. There is a significant literature discussing the implications of biometric technology use, though much of this work is theoretical in nature and there is comparatively little empirically grounded work with a focus on the biometric user experience. This thesis presents research investigating biometric authentication from a user-centric perspective. The principal aims of this research were to deepen our understanding of the usability and acceptability of biometric authentication and use this knowledge to improve design. A series of controlled evaluations are presented, where biometric systems and different aspects of system design were investigated. To understand wider implementation issues, a field trial of a biometric system in a real-world environment was also carried out. A second strand of research focused on how biometrics are perceived and both survey and interview approaches were used to explore this issue. In general the empirical work can be characterized by a trend of structured, quantitative methodologies leading into less-structured approaches as contextual and experiential aspects of system use were investigated.A framework for the biometric user experience is presented based on this work. The framework is used to structure the design guidelines and knowledge emerging from this work. A methodology for the user-centric evaluation of biometrics is also proposed. The results from this project further our understanding of usable system design, but biometrics have proven to be an emotive technology and implementation remains a complex issue