Multimodal biometric system using particle swarm based feature selection

Abstract

A scientific discipline which includes methods that enable identification of people through their behavioral or physical characteristics or sometimes through both is known as Biometrics. Enhancement of security and the increase in its demand has paved way to an interest in an automated method of personal authentication which is based on Biometrics in recent times. Owing to the availability of multiple factors of evidence biometric systems that are multi-modal are considered to be more reliable. In visual interactions among humans, facial recognition is one of the most commonly acceptable biometrics. Selection of features is a challenging task as the interaction among the facial features can be complex. Therefore, Particle Swarm Optimization (PSO) which is a feature selection based on algorithm has been proposed as suitable for the multi-modal biometric system which in turn can bring down significantly the dimension by making use of different fusion schemes and identifying optimal features from among them.