Person identification using face and iris multimodal biometric system

Abstract

Multimodal biometric systems fuse information using more than one physical and/or behavioral characteristics of a person to improve the recognition accuracy for person identification and alleviate the single biometric trait limitations. Focus of this paper is on combining the strengths of face and iris modalities to obtain better recognition accuracy for person identification by using several feature extractors, score normalization and fusion techniques. Face and iris features are extracted separately using global and local feature extractors and then the fusion of these modalities is performed. The experiments are conducted on ORL face database and CASIA iris database. Evaluation of experimental results demonstrates the enhancement of recognition for face-iris fusion compared to the individual modalities.