Toggle Main Menu Toggle Search

Open Access padlockePrints

Accent Issues in Large Vocabulary Continuous Speech Recognition

Lookup NU author(s): Tao Chen

Downloads

Full text for this publication is not currently held within this repository. Alternative links are provided below where available.


Abstract

This paper addresses accent1 issues in large vocabulary continuous speech recognition. Cross-accent experiments show that the accent problem is very dominant in speech recognition. Analysis based on multivariate statistical tools (principal component analysis and independent component analysis) confirms that accent is one of the key factors in speaker variability. Considering different applications, we proposed two methods for accent adaptation. When a certain amount of adaptation data was available, pronunciation dictionary modeling was adopted to reduce recognition errors caused by pronunciation mistakes. When a large corpus was collected for each accent type, accent-dependent models were trained and a Gaussian mixture model-based accent identification system was developed for model selection. We report experimental results for the two schemes and verify their efficiency in each situation.


Publication metadata

Author(s): Huang C, Chen T, Chang E

Publication type: Article

Publication status: Published

Journal: International Journal of Speech Technology

Year: 2004

Volume: 7

Issue: 2-3

Pages: 141-153

Print publication date: 01/04/2004

ISSN (print): 1381-2416

ISSN (electronic): 1572-8110

Publisher: Springer New York LLC

URL: http://dx.doi.org/10.1023/B:IJST.0000017014.52972.1d

DOI: 10.1023/B:IJST.0000017014.52972.1d


Altmetrics

Altmetrics provided by Altmetric


Share