No Cover Image

Conference Paper/Proceeding/Abstract 221 views

Continuous speech recognition using syllables

Rhys Jones Orcid Logo, John Mason, Simon Downey

Eurospeech '97, Pages: 1171 - 1174

Swansea University Authors: Rhys Jones Orcid Logo, John Mason

Full text not available from this repository: check for access using links below.

Abstract

The vast majority of work in continuous speech recognition uses phoneme-like units as the basic recognition component. The work presented here investigates the practicability of syllable-like units as the building blocks for recognition. A phonetically annotated telephony database is analysed at the...

Full description

Published in: Eurospeech '97
ISSN: 1018-4074
Published: Grenoble, France European Speech Communication Association: ESCA 1997
Online Access: Check full text

URI: https://cronfa.swan.ac.uk/Record/cronfa63336
Tags: Add Tag
No Tags, Be the first to tag this record!
Abstract: The vast majority of work in continuous speech recognition uses phoneme-like units as the basic recognition component. The work presented here investigates the practicability of syllable-like units as the building blocks for recognition. A phonetically annotated telephony database is analysed at the syllable level, and a set of syllable-based Hidden Markov Models (HMMs) are built. Refinements including the introduction of syllable-level bigram probabilities, word- and syllable-level insertion penalties, and the investigation of different model topologies are found to improve recogniser performance. It is found that the syllable-based recogniser gives recognition accuracies of over 60%, which compares with 35% as the baseline accuracy for monophone recognition. It is envisaged that practical applications of syllable recognition could be in a hybrid system, where the most common syllable HMMs would be used in conjunction with whole-word and phoneme models.
College: Faculty of Humanities and Social Sciences
Start Page: 1171
End Page: 1174