DECISION TREE-BASED TRIPHONES ARE ROBUST AND PRACTICAL FOR MANDARIAN SPEECH RECOGNITION

Yi Liu; Pascale Fung

DECISION TREE-BASED TRIPHONES ARE ROBUST AND PRACTICAL FOR MANDARIAN SPEECH RECOGNITION

Yi Liu, Pascale Fung

Department of Electronic and Computer Engineering

Research output: Contribution to conference › Conference Paper › peer-review

4 Citations (Scopus)

Abstract

In large-vocabulary, speaker-independent speech recognition systems, modeling of vocabulary words by subword units is mandatory. This paper studies the use of triphone units for Mandarin speech recognition compared to biphone and context-independent phonetic units. In order to solve unseen triphones in speech recognition, decision-tree based clustering is used in triphone units. This method achieves high recognition performance with limited training data and also reduces the model training time. The robustness and effectiveness of the cross-word, tree-based triphone units have been proved by the speaker-independent continuous Mandarin speech recognition task. The training computation time reduces by about 2.3 times after tying states for triphone models, the recognition syllable accuracy increases 28.7% compared to monophone units and by 13.5% compared to biphone units.

Original language	English
Pages	895-898
Number of pages	4
Publication status	Published - 1999
Event	6th European Conference on Speech Communication and Technology, EUROSPEECH 1999 - Budapest, Hungary Duration: 5 Sept 1999 → 9 Sept 1999

Conference

Conference	6th European Conference on Speech Communication and Technology, EUROSPEECH 1999
Country/Territory	Hungary
City	Budapest
Period	5/09/99 → 9/09/99

Bibliographical note

Publisher Copyright:
© 1999 6th European Conference on Speech Communication and Technology, EUROSPEECH 1999. All rights reserved.

Cite this

@conference{b50ee730457941dd968c7b7225c41583,

title = "DECISION TREE-BASED TRIPHONES ARE ROBUST AND PRACTICAL FOR MANDARIAN SPEECH RECOGNITION",

abstract = "In large-vocabulary, speaker-independent speech recognition systems, modeling of vocabulary words by subword units is mandatory. This paper studies the use of triphone units for Mandarin speech recognition compared to biphone and context-independent phonetic units. In order to solve unseen triphones in speech recognition, decision-tree based clustering is used in triphone units. This method achieves high recognition performance with limited training data and also reduces the model training time. The robustness and effectiveness of the cross-word, tree-based triphone units have been proved by the speaker-independent continuous Mandarin speech recognition task. The training computation time reduces by about 2.3 times after tying states for triphone models, the recognition syllable accuracy increases 28.7\% compared to monophone units and by 13.5\% compared to biphone units.",

author = "Yi Liu and Pascale Fung",

note = "Publisher Copyright: {\textcopyright} 1999 6th European Conference on Speech Communication and Technology, EUROSPEECH 1999. All rights reserved.; 6th European Conference on Speech Communication and Technology, EUROSPEECH 1999 ; Conference date: 05-09-1999 Through 09-09-1999",

year = "1999",

language = "English",

pages = "895--898",

}

TY - CONF

T1 - DECISION TREE-BASED TRIPHONES ARE ROBUST AND PRACTICAL FOR MANDARIAN SPEECH RECOGNITION

AU - Liu, Yi

AU - Fung, Pascale

PY - 1999

Y1 - 1999

N2 - In large-vocabulary, speaker-independent speech recognition systems, modeling of vocabulary words by subword units is mandatory. This paper studies the use of triphone units for Mandarin speech recognition compared to biphone and context-independent phonetic units. In order to solve unseen triphones in speech recognition, decision-tree based clustering is used in triphone units. This method achieves high recognition performance with limited training data and also reduces the model training time. The robustness and effectiveness of the cross-word, tree-based triphone units have been proved by the speaker-independent continuous Mandarin speech recognition task. The training computation time reduces by about 2.3 times after tying states for triphone models, the recognition syllable accuracy increases 28.7% compared to monophone units and by 13.5% compared to biphone units.

AB - In large-vocabulary, speaker-independent speech recognition systems, modeling of vocabulary words by subword units is mandatory. This paper studies the use of triphone units for Mandarin speech recognition compared to biphone and context-independent phonetic units. In order to solve unseen triphones in speech recognition, decision-tree based clustering is used in triphone units. This method achieves high recognition performance with limited training data and also reduces the model training time. The robustness and effectiveness of the cross-word, tree-based triphone units have been proved by the speaker-independent continuous Mandarin speech recognition task. The training computation time reduces by about 2.3 times after tying states for triphone models, the recognition syllable accuracy increases 28.7% compared to monophone units and by 13.5% compared to biphone units.

UR - https://www.scopus.com/pages/publications/0141901978

M3 - Conference Paper

AN - SCOPUS:0141901978

SP - 895

EP - 898

T2 - 6th European Conference on Speech Communication and Technology, EUROSPEECH 1999

Y2 - 5 September 1999 through 9 September 1999

ER -

DECISION TREE-BASED TRIPHONES ARE ROBUST AND PRACTICAL FOR MANDARIAN SPEECH RECOGNITION

Abstract

Conference

Bibliographical note

Other files and links

Fingerprint

Cite this