In this paper we investigate benefits of classifier combination for a multimodal system for personal identity verification. The system uses frontal face images and speech. We show that a sophisticated fusion strategy enables the system to outperform its fa ...
IDIAP1998