Score-Level Fusion of Fingerprint, Face and Iris based on Choquet Integral

Abstract

Multimodal biometric systems fuse information from multiple modalities to overcome limitations of the unimodal biometric system. In this paper, we highlight recent biometric technologies that have been applied across a variety of fields. Furthermore, we propose the fusion of three unimodal biometric verification systems, based on the fingerprint, the face and the iris. The fusion of modalities is made at score level based on Choquet integral using particle swarm optimization to set their fuzzy measures. The experimental results show that the proposal contributes to improve the unimodal performance system leading to a low equal error rate of 0.1%.