Multimodal biometric system based on fusion techniques: a review

Abstract

ABSTRACT Biometric allude to an automatic procedure of acknowledging an individual utilizing their behavioral or physical characteristics. The biometric framework, which utilizes one cue for authentication is termed as unimodal biometric. The unimodal biometric framework confronts numerous snags like dearth of distinctiveness, universality intra-class similarity and multimodal biometric is one of best options to conquer these issues, which is a framework that utilizes two or more cues for authentication. This study presents the overview of multimodal biometric recognition systems. Multimodal biometric recognition systems augment the security and concealment of digital information. From last two decades, there are a lot of research work on information fusion. We have discussed recent trends in multimodal biometric depending upon the type of fusion scheme and the level of fusion i.e. sensor level or feature level fusion, decision level fusion, score level fusion and hybrid fusion level. The types of fusion are conversed in detail with their individual merits and demerits. In addition to that, the methodologies, employed databases and accuracy results of the existing works are presented to showcase the profound usage of multimodal biometric design. The paper is targeted toward presenting a comprehensive review of different fusion schemes in combining various biometric modalities.