Development of Multimodal Biometric Recognition System Using Hybrid Fusion Techniques for Efficient Authentication

Abstract

newlineIn recent years the upsurge in the advents of contemporary technical assets results in augmented prerequisite of precise and robust security systems. The utmost prevailing alternative for the security requirements is deployment of effective biometric authentication frameworks. The term quotbiometricquot refers to a method of automatically identifying people based on their behavioural or physical physiognomies like palm-prints, fingerprints, voice, finger vein, ECG, finger-knuckle-prints or iris. These physiognomies are stated as biometric cues, modalities or traits. However, depending upon the kind of application where the framework is to be utilized, cues can be chosen. A general biometric recognition framework comprises of four sections such as pre-processing of images, extraction of features, classification and decision stage. Unimodal biometric refers to a biometric framework that only requires one cue for authentication. Numerous challenges, including a lack of distinctiveness, non-universality, and intra-class similarity, plague the unimodal biometric framework due to which it lags in precision and one of the greatest solutions to these problems is multimodal biometric, which is an authentication framework that makes use of two or more traits. In this concern, the multimodal biometrics has gained more consideration of investigators as it utilizes the proven and effective strategy to integrate the multiple human traits for authentication.