A COMPREHENSIVE FRAMEWORK FOR VOICE BIOMETRICS

Abstract

Recognizing someone by his/her voice is called Automatic Speaker Recognition (ASR). Biometric protection comes under Speaker ID biometric is concerned with individual characteristics Voiceprint or biometric authentication is used to identify a person. It depends on an individual's voice activity or physiology. Behavioral biometrics includes speech, signature, keystroke, and typing while physiological biometrics includes face, retina, fingerprints, and DNA now biometric authentication is cutting-edge.Several forms of speech recognition and feature extraction systems were employed in this paper. A novel feature extraction method was proposed using Sub-band based SBC, and energy calculation was also covered. For a specific combination of 350 users, our approach achieves recognition rate above 94% with EER less than 6%. The output is measured on two separate sets of data, clean and degraded.