IDENTITY AUTHENTICATION USING VOICE BIOMETRICS TECHNIQUE

U A Kamalu, A. Raji, V.I. Nnebedum

Abstract: Identification of people using name, appearances, badges, tags and register may be effective may be in a small organization. However, as the size of the organization or society increases, these simple ways of identifying individual become ineffective. Therefore, it may be necessary to employ additional and more sophisticated means of authenticating the identity of people as the population increases. Voice Biometrics is a method by which individuals can be uniquely identified by evaluating one or more distinguishing biological traits associated with the voice of such individuals. In this paper, an unconstrained text-independent recognition system using the Gaussian Mixture Model was applied to match recorded voice to stored voice for the purpose of identification of individual. Recorded voices were processed and stored in the enrollment phase while probing voices were used for comparison in the verification/recognition phase of the system.

Keywords: Model, Biometric, verification, enrollment, database, authentication, matching, identity.

DOI: https://doi.org/10.15623/ijret.2015.0412025