PHONEME RECOGNITION IN VARIOUS STYLES OF UTTERANCE BASED ON MUTUAL INFORMATION CRITERION

Shigeki Okawa; Tetsunori Kobayashi; Katsuhiko Shirai

PHONEME RECOGNITION IN VARIOUS STYLES OF UTTERANCE BASED ON MUTUAL INFORMATION CRITERION

Shigeki Okawa, Tetsunori Kobayashi, Katsuhiko Shirai

School of Fundamental Science and Engineering

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

3 Citations (Scopus)

Abstract

This paper discusses a highly reliable phoneme recognition method in various styles of utterance based on mutual information criterion. Mutual information is a good measure to build an effective phoneme dictionary in the process of optimal selection of acoustic features and integration of clusters. Using VQ code sequences organized by the hierarchical clustering method, phonemic likelihoods for each frame can be calculated. Phoneme recognition is performed with applying phonemic duration and bigram constraints of phonemes. Also, we cover an iterative training mechanism of the phoneme dictionary. The correct rate for phoneme is improved to 90.5% (8.4% insertion, 7.0% deletion) in the speaker independent recognition experiment for the continuous utterance.

Original language	English
Pages	1911-1914
Number of pages	4
Publication status	Published - 1994
Event	3rd International Conference on Spoken Language Processing, ICSLP 1994 - Yokohama, Japan Duration: 1994 Sept 18 → 1994 Sept 22

Conference

Conference	3rd International Conference on Spoken Language Processing, ICSLP 1994
Country/Territory	Japan
City	Yokohama
Period	94/9/18 → 94/9/22

ASJC Scopus subject areas

Language and Linguistics
Linguistics and Language

Cite this

@conference{e59feec2e8ea4e629398dd668b258719,

title = "PHONEME RECOGNITION IN VARIOUS STYLES OF UTTERANCE BASED ON MUTUAL INFORMATION CRITERION",

abstract = "This paper discusses a highly reliable phoneme recognition method in various styles of utterance based on mutual information criterion. Mutual information is a good measure to build an effective phoneme dictionary in the process of optimal selection of acoustic features and integration of clusters. Using VQ code sequences organized by the hierarchical clustering method, phonemic likelihoods for each frame can be calculated. Phoneme recognition is performed with applying phonemic duration and bigram constraints of phonemes. Also, we cover an iterative training mechanism of the phoneme dictionary. The correct rate for phoneme is improved to 90.5% (8.4% insertion, 7.0% deletion) in the speaker independent recognition experiment for the continuous utterance.",

author = "Shigeki Okawa and Tetsunori Kobayashi and Katsuhiko Shirai",

note = "Funding Information: The authors are grateful to the members of Laboratory for Spoken Language Processing of Waseda University for their help and discussions. This work is partly supported by the Grant-in-Aid for Scientific Research from the Ministry Education, Science and Culture of Japan, No. 05241103. Publisher Copyright: {\textcopyright} 1994 3rd International Conference on Spoken Language Processing, ICSLP 1994. All rights reserved.; 3rd International Conference on Spoken Language Processing, ICSLP 1994 ; Conference date: 18-09-1994 Through 22-09-1994",

year = "1994",

language = "English",

pages = "1911--1914",

}

TY - CONF

T1 - PHONEME RECOGNITION IN VARIOUS STYLES OF UTTERANCE BASED ON MUTUAL INFORMATION CRITERION

AU - Okawa, Shigeki

AU - Kobayashi, Tetsunori

AU - Shirai, Katsuhiko

N1 - Funding Information: The authors are grateful to the members of Laboratory for Spoken Language Processing of Waseda University for their help and discussions. This work is partly supported by the Grant-in-Aid for Scientific Research from the Ministry Education, Science and Culture of Japan, No. 05241103. Publisher Copyright: © 1994 3rd International Conference on Spoken Language Processing, ICSLP 1994. All rights reserved.

PY - 1994

Y1 - 1994

N2 - This paper discusses a highly reliable phoneme recognition method in various styles of utterance based on mutual information criterion. Mutual information is a good measure to build an effective phoneme dictionary in the process of optimal selection of acoustic features and integration of clusters. Using VQ code sequences organized by the hierarchical clustering method, phonemic likelihoods for each frame can be calculated. Phoneme recognition is performed with applying phonemic duration and bigram constraints of phonemes. Also, we cover an iterative training mechanism of the phoneme dictionary. The correct rate for phoneme is improved to 90.5% (8.4% insertion, 7.0% deletion) in the speaker independent recognition experiment for the continuous utterance.

AB - This paper discusses a highly reliable phoneme recognition method in various styles of utterance based on mutual information criterion. Mutual information is a good measure to build an effective phoneme dictionary in the process of optimal selection of acoustic features and integration of clusters. Using VQ code sequences organized by the hierarchical clustering method, phonemic likelihoods for each frame can be calculated. Phoneme recognition is performed with applying phonemic duration and bigram constraints of phonemes. Also, we cover an iterative training mechanism of the phoneme dictionary. The correct rate for phoneme is improved to 90.5% (8.4% insertion, 7.0% deletion) in the speaker independent recognition experiment for the continuous utterance.

UR - http://www.scopus.com/inward/record.url?scp=85009222101&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85009222101&partnerID=8YFLogxK

M3 - Paper

AN - SCOPUS:85009222101

SP - 1911

EP - 1914

T2 - 3rd International Conference on Spoken Language Processing, ICSLP 1994

Y2 - 18 September 1994 through 22 September 1994

ER -

PHONEME RECOGNITION IN VARIOUS STYLES OF UTTERANCE BASED ON MUTUAL INFORMATION CRITERION

Abstract

Conference

ASJC Scopus subject areas

Other files and links

Fingerprint

Cite this