Automatic speaker recognition using deep neural network classifiers

Abdikarim Ali Moumin, Smitha S. Kumar

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

The advances in modern computing technologies have achieved a breakthrough in the fields of artificial intelligence (AI) and the Internet of Things (IoT). One of the major achievements in the recent history is the ability of the computer software to classify and recognize some of the objects or sounds by learning data. In this paper, we have trained the software to recognize people using their voice utterances using TIMIT Acoustic Phonetic Continuous Speech Corpus. The speaker identity is enrolled by acquiring voice samples of the speaker. Relevant features are extracted, and a model is built using the extracted feature vectors. A pattern matching classification is applied to the model using artificial neural network techniques. Speaker verification system is built using Kaldi libraries to analyze acoustic features, while x-vector training is implemented using Tensor Flow. To achieve better performance, we have implemented a combination of multiple layers of TDNN (Time Delay Neural Networks) and LSTM (Long Short-Term Memory) deep neural networks.

Original languageEnglish
Title of host publication2021 2nd International Conference on Computation, Automation and Knowledge Management (ICCAKM)
PublisherIEEE
Pages282-286
Number of pages5
ISBN (Electronic)9781728194912
DOIs
Publication statusPublished - 25 Feb 2021
Event2nd International Conference on Computation, Automation and Knowledge Management 2021 - Dubai, United Arab Emirates
Duration: 19 Jan 202121 Jan 2021

Conference

Conference2nd International Conference on Computation, Automation and Knowledge Management 2021
Abbreviated titleICCAKM 2021
Country/TerritoryUnited Arab Emirates
CityDubai
Period19/01/2121/01/21

Keywords

  • DNN
  • TDNN and LSTM

ASJC Scopus subject areas

  • Computer Networks and Communications
  • Computer Science Applications
  • Information Systems and Management
  • Control and Optimization

Fingerprint

Dive into the research topics of 'Automatic speaker recognition using deep neural network classifiers'. Together they form a unique fingerprint.

Cite this