Acoustic model selection using limited data for accent robust speech recognition

Maryam Najafian, Saeid Safavi, Abualsoud Hanani, Martin Russell

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

8 Citations (Scopus)
197 Downloads (Pure)

Abstract

This paper investigates techniques to compensate for the effects of regional accents of British English on automatic speech recognition (ASR) performance. Given a small amount of speech from a new speaker, is it better to apply speaker adaptation, or to use accent identification (AID) to identify the speaker's accent followed by accent-dependent ASR? Three approaches to accent-dependent modelling are investigated: using the `correct' accent model, choosing a model using supervised (ACCDIST-based) accent identification (AID), and building a model using data from neighbouring speakers in `AID space'. All of the methods outperform the accent-independent model, with relative reductions in ASR error rate of up to 44%. Using on average 43s of speech to identify an appropriate accent-dependent model outperforms using it for supervised speaker-adaptation, by 7%.
Original languageEnglish
Title of host publication2014 Proceedings of the 22nd European Signal Processing Conference (EUSIPCO)
PublisherInstitute of Electrical and Electronics Engineers (IEEE)
Pages1786 - 1790
Number of pages5
ISBN (Print)978-0-9928626-1-9
Publication statusPublished - 2014
Event22nd European Signal Processing Conference, EUSIPCO 2014 - Lisbon, United Kingdom
Duration: 1 Sep 20145 Sep 2014

Conference

Conference22nd European Signal Processing Conference, EUSIPCO 2014
Country/TerritoryUnited Kingdom
CityLisbon
Period1/09/145/09/14

Keywords

  • speech recognition
  • accent recognition

Fingerprint

Dive into the research topics of 'Acoustic model selection using limited data for accent robust speech recognition'. Together they form a unique fingerprint.

Cite this