Brief Review of Short Utterance Speaker Verification Systems

Abstract

Due to technological improvements many methods have been proposed for speaker verification. While performance is satisfactory with large amounts of speech data, there is significant degradation in performance with short utterances. Many research works have been carried out to handle short utterance issue of the speaker verification systems used in real-world scenario. In this paper we primarily emphasis on the survey of different feature extraction methods for textindependent speaker verification. We first briefly review conventional systems to show its progress. In this work, we present a brief review of features that are used to capture speaker information at different analysis lengths of speech utterance. We also put the major findings and challenges of this research feview in a nutshell.