Multimodel biometric system: Fusion techniques and their comparison

Abstract

Security and surveillance related technologies are rapidly growing day by day in the modern world to provide the security to the person and to the property. The biometric system is the only one that can provide accurate, secure and reliable scheme for personal verification. Biometric systems are broadly categories into two types: Uni-modal and Multi-modal. By comparing both these systems, it is clear that multi-modal biometric systems are gathering more popularity as it has the capability of dealing challenges involved in uni-modal biometric systems such as: non-universality, captured data noise, variations between same class, spoof attacks and unacceptable error rates. This paper will discuss various levels of information fusion, fusion techniques used at different levels and comparison among these fusion levels.