Comparison of levels and fusion approaches for multimodal biometrics

Abstract

The biometric-based authentication system occupies maximal space in the field of security administration. Biometric applications are swiftly accelerating in day-to-day life such as computer login, smart homes, online banking, hospitals, border areas, industries, forensics, e-voting attendance system and investigation of crime. A reliable and accurate recognition body can be achieved with multimodal biometric methodologies. In this paper, we discuss starting with an introduction to biometric systems followed by their classification, and advantages as well as disadvantages. In today’s world, most of the systems are unimodal biometrics having a lot of limitations to overcome those multimodal biometrics comes in to picture. In this paper we have discussed comprehensive representation on the system of multimodal biometric, various modes of undertakings, the significance of information fusion, a different section is allotted on the various possible levels of fusion involving sensor-level, feature-level, score-level, and decision -level as well as different rules of fusion.