Towards optimised implementations of multimodal biometric configurations

Multimodal biometric solutions can provide increased reliability and robustness in many applications where identity verification is important, and especially so in highly interactive scenarios such as those represented by internet based transactions. Yet multimodal configurations increase complexity and raise questions about the need for effective system optimisation. This paper describes an approach that uses genetic algorithms to evolve efficient configurations in such applications, and shows how such structures can be embedded within a broader implementation framework, which is especially suited to supporting typical transactions.