Machine learning for voice recognition

Kinkiri, Saritha; Melis, Wim J.C.; Keates, Simeon

Abstract

Verbal communication is very important to us humans, but using thisperforming verbal communication to communicateion with machines still faces particular challenges. Therefore, researchers are trying to find ways to make communication with a machine more similar to communicating with other people, for which two systems have been identified: speech and voice recognition. While speech recognition has aimed to become speaker independent, voice recognition focuses on identifying the speaker, by looking at the tone of the voice, which is affected by the physical characteristics of that person. This requires one to identify these unique tonal features, to then train a system with this data. Being able to perform this identification well, would also bring benefit to speech recognition by allowing the system to adjust to the characteristics of that speaker and how he/she produces their sounds.

Citation

Kinkiri, S., Melis, W. J., & Keates, S. (2017, June). Machine learning for voice recognition. Presented at 2nd Medway Engineering Converence on Systems 2017, University of Greenwich, United Kingdom

• Presentation Conference Type: Conference Paper (published)
• Conference Name: 2nd Medway Engineering Converence on Systems 2017
• Start Date: Jun 6, 2017
• End Date: Jun 6, 2017
• Acceptance Date: May 22, 2017
• Publication Date: 2017
• Deposit Date: Feb 7, 2019
• Series ISSN: 2398-306X
• Book Title: The Second Medway Engineering Conference on Systems: Efficiency, Sustainability and Modelling
• Keywords: Machine learning, Communication, Voice recognition, Speech recognition, Security, Biometric authentication
• Public URL: http://researchrepository.napier.ac.uk/Output/1497086