FUSION OF MULTIMODAL UNCONSTRAINED BIOMETRIC AUTHENTICATION USING MACHINE LEARNING TECHNIQUE

Abstract

The term "biometric identification" pertains to the use of physiological and behavioural attributes for the purpose of distinguishing and verifying the identity of a person. The reliability of biometric identification surpasses that of older techniques. This paper presents a comprehensive review of the identification process, including a concise explanation of the functioning of biometric systems, the rationale behind its efficacy as a viable substitute for conventional identification methods, and the evaluation of biometric systems in terms of their performance. In this work, a novel bag of classifiers approach is proposed. Accuracy, specificity, sensitivity, F1-score, MCC, and AUC measurements are used for performance comparison. The performance of the proposed method is found to be considerable over the other methods.

Keyword: Biometric Identification, Feature Selection, Feature Fusion, Machine Learning, Multimodal, Biometric Authentication