Improving ASR performance using context‐dependent phoneme models
| Published date | 02 February 2010 |
| Date | 02 February 2010 |
| DOI | https://doi.org/10.1108/13287261011032652 |
| Pages | 56-69 |
| Author | Husniza Husni,Zulikha Jamaludin |
| Subject Matter | Information & knowledge management |
JSIT
12,1
56
Journal of Systems and Information
Technology
Vol. 12 No. 1, 2010
pp. 56-69
#Emerald Group Publishing Limited
1328-7265
DOI 10.1108/13287261011032652
Improving ASR performance
using context-dependent
phoneme models
Husniza Husni and Zulikha Jamaludin
UUM College of Arts and Sciences, Universiti Utara Malaysia,
Sintok, Malaysia
Abstract
Purpose – The purpose of this paper is to present evidence of the need to have a carefully designed
lexical model for speech recognition for dyslexic children reading in Bahasa Melayu (BM).
Design/methodology/approach – Data collection is performed to obtain the most frequent reading
error patterns and the reading recordings. Design and development of the lexical model considers the
errors for better recognition accuracy.
Findings – It is found that the recognition accuracy is increased to 75 percent when using context-
dependent (CD) phoneme model and phoneme refinement rule. Comparison between context-
independent phoneme models and CD phoneme model is also presented.
Research limitations/implications – The most frequent errors recognized and obtained from data
collection and analysis illustrate and support that phonological deficit is the major facto r for reading
disabilities in dyslexics.
Practical implications – This paper provides the first step towards materializing an automated
speech recognition (ASR)-based application to support reading for BM, which is the first language in
Malaysia.
Originality/value – The paper contributes to the knowledge of the most frequent error patterns for
dyslexic children’s reading in BM and to the knowledge that a CD phoneme model together with the
phonemerefinement rule canbuilt up a more fine-tuned lexicalmodel for an ASR specifically fordyslexic
children’s reading isolated words in BM.
Keywords Speech recognition equipment, Reading, Dyslexia, Children (age groups), Malaysia
Paper type Research paper
1. Introduction
The demand for automatedspeech recognition(ASR) technology to help children to read
has increased significantly due to the potential that ASR has (Steidl et al.,2003; Raskind
and Higgins, 1999; Higgins and Raskind, 2000). Such technology has been seen as an
alternative way of teaching reading to children. In fact, ASR is the key towards an
automatic readingtutor where it is used to ‘‘listen’’to the readings, track the reading,and
detect miscues(Mostow et al., 1994; Russell et al., 1996;Hagen et al., 2004; Nix et al.,1998;
Williams et al., 2000; Duchateau et al.,2006; Li et al., 2007, 2008; Liu et al.,2008).
With the advancement of the ASR technology in teaching and training children to
read, its potential could be manipulated to provide help for children especially those
with dyslexia. Dyslexia is a condition that impedes phonological awareness, which is
strongly related to reading ability especially in the letter-sound cor respondence area.
Despite reading, dyslexia also causes problems in other skills suc h as writing, spelling,
and motor skills as well as memory and cognition. In favor of the fact that reading is
the key towards knowledge acquisition, help should be provided to these children from
The current issue and full text archive of this journal is available at
www.emeraldinsight.com/1328-7265.htm
Thank you to special education teachers (Bahasa Melayu) of SK Taman TunDr Ismail (2), Kuala
Lumpur, and SK Jalan Datuk Kumbar, Alor Setar for the positive cooperation received
throughout the entire data collection period. A special thank you also goes to Assistant
Professor John-Paul Hosom of CSLU, OGI, USA, for his consistent help.
To continue reading
Request your trial