|
Advanced search
Previous page
|
Title
Comparing grapheme-based and phoneme-based speech recognition for Afrikaans |
Full text
http://www.prasa.org/index.php/2012-03-07-10-55-15; http://hdl.handle.net/10204/6492 |
Date
2012 |
Author(s)
Basson, WD; Davel, MH |
Abstract
This paper compares the recognition accuracy of a phoneme-based automatic speech recognition system with that of a grapheme-based system, using Afrikaans as case study. The first system is developed using a conventional pronunciation dictionary, while the latter system uses the letters of each word directly as the acoustic units to be modelled. We ensure that the pronunciation dictionary we use is highly accurate and then investigate the extent to which ASR performance degrades when the dictionary is removed.We analyse this effect at different data set sizes and classify the causes of performance degradation. With grapheme-based ASR outperforming phoneme-based ASR in certain word categories, we find that relative error rates are highly dependent on word category, which points towards strategies for compensating for grapheme-based inaccuracies. - PRASA 2012, CSIR International Convention Centre, Pretoria, 29-30 November 2012 |
Subject(s)
Automatic speech recognition; ASR; Grapheme-based system |
Language
en |
Publisher
PRASA 2012 |
Relation
Workflow;10205 |
Type of publication
Presentation |
Identifier
Basson, W.D and Davel, M.H. 2012. Comparing grapheme-based and phoneme-based speech recognition for Afrikaans. In: PRASA 2012, CSIR International Convention Centre, Pretoria, 29-30 November 2012; 978-0-620-54601-0 |
Repository
Pretoria - Council for Scientific and Industrial Research (CSIR)
|
Added to C-A: 2017-01-27;13:46:11 |
© Connecting-Africa 2004-2024 | Last update: Friday, March 8, 2024 |
Webmaster
|