Intelligent management of multimodal biometric transactions

Abstract

Multimodal biometric systems provide both increased reliability and greater flexibility to users, yet also increase the degree of complexity in implementation and system management, especially where access permissions, security levels, data categorizations and user characteristics show a high degree of diversity. Also, sources of unreliability in biometric authentication are often traceable to problems of enrolment, and especially to poor data acquisition, and hence introducing a greater degree of monitoring and increasing intelligent user guidance at the system interface can make a significant impact on the overall performance of an identity authentication system, especially in a multimodal scenario. This paper illustrates some of these problems in practice and outlines an implementation framework which reconciles the need to support more effective interaction mechanisms with the wider problems of managing the complexity of data flow which arises in such systems.