Skip to content
Publications

Article · November 7, 2021

СИСТЕМАТИЧЕСКИЙ ОБЗОР И АНАЛИЗ ОСОБЕННОСТЕЙ ИДЕНТИФИКАЦИИ ПО ГОЛОСУ

О. Ж. Мамырбаев, Aizat Kydyrbekova, А. Т. Ахмедиярова, Mussa Turdalyuly, Nurbapa Mekebayev

Abstract

Voice identification is the process of identifying a speaker by a given utterance by comparing voice biometrics of a utterance with those utterance models that were saved in advance. Voice identification technologies have gained a new direction due to advances in artificial intelligence and are widely used in various fields. Character extraction is one of the most important aspects of voice identification, which significantly affects the identification process and performance. This systematic review is conducted to identify, compare and analyze various approaches, methods and algorithms for extracting features for voice identification, to provide background information on character retrieval approaches for voice identification applications and future research. The study examined models: based on patterns, based on vector quantization, dynamic time transfer, histogram model, stochastic models, Gaussian mixture models and hidden Markov model, based on Mel-frequency cepstral coefficients, generative or vector quantization, discriminatory models (usually using machine learning methods such as SVM and ANN).This study showed that the current trend of identification research is to develop a robust, universal voice identification structure for solving important voice identification problems, such as adaptability, complexity, multilingual recognition, and resistance to noise. The results presented in this study are based on past publications, quotations and the number of implementations, the quotes being the most relevant. This article also presents the general process of voice identification.

Speech Recognition and SynthesisSpeech and Audio ProcessingBiometric Identification and Securityspeaker identificationvoice biometricsfeature extraction

More by Aizat Kydyrbekova