Stress detection based on multi-class probabilistic support vector machines for accented English speech

Jhing Fa Wang; Gung Ming Chang; Jia Ching Wang; Shun Chieh Lin

doi:10.1109/CSIE.2009.739

Stress detection based on multi-class probabilistic support vector machines for accented English speech

Jhing Fa Wang, Gung Ming Chang, Jia Ching Wang, Shun Chieh Lin

Department of Computer Science and Information Engineering

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

6 Scopus citations

Abstract

A stress detection based on multi-class probabilistic support vector machines (MCP-SVMs) is proposed for classifying speech into following categories - no stress, primary stress, and secondary stress. The stress classifier is performed with a feature set including perceptual features, MFCC, delta-MFCC and delta-delta-MFCC. To observe that speakers from the same accent regions had similar tendencies in mispronunciations including word stress, this work uses English Across Taiwan (EAT) to represent Taiwanese-accented English speech corpora. The overall performance in the experimental results achieves about 84% classification of accuracy.

Original language	English
Title of host publication	2009 WRI World Congress on Computer Science and Information Engineering, CSIE 2009
Pages	346-350
Number of pages	5
DOIs	https://doi.org/10.1109/CSIE.2009.739
State	Published - 2009
Event	2009 WRI World Congress on Computer Science and Information Engineering, CSIE 2009 - Los Angeles, CA, United States Duration: 31 Mar 2009 → 2 Apr 2009

Publication series

Name	2009 WRI World Congress on Computer Science and Information Engineering, CSIE 2009
Volume	7

Conference

Conference	2009 WRI World Congress on Computer Science and Information Engineering, CSIE 2009
Country/Territory	United States
City	Los Angeles, CA
Period	31/03/09 → 2/04/09

Access to Document

10.1109/CSIE.2009.739

Cite this

Wang, J. F., Chang, G. M., Wang, J. C., & Lin, S. C. (2009). Stress detection based on multi-class probabilistic support vector machines for accented English speech. In 2009 WRI World Congress on Computer Science and Information Engineering, CSIE 2009 (pp. 346-350). Article 5170340 (2009 WRI World Congress on Computer Science and Information Engineering, CSIE 2009; Vol. 7). https://doi.org/10.1109/CSIE.2009.739

@inproceedings{99767da2841f43d09e54a74700f9c737,

title = "Stress detection based on multi-class probabilistic support vector machines for accented English speech",

abstract = "A stress detection based on multi-class probabilistic support vector machines (MCP-SVMs) is proposed for classifying speech into following categories - no stress, primary stress, and secondary stress. The stress classifier is performed with a feature set including perceptual features, MFCC, delta-MFCC and delta-delta-MFCC. To observe that speakers from the same accent regions had similar tendencies in mispronunciations including word stress, this work uses English Across Taiwan (EAT) to represent Taiwanese-accented English speech corpora. The overall performance in the experimental results achieves about 84% classification of accuracy.",

author = "Wang, {Jhing Fa} and Chang, {Gung Ming} and Wang, {Jia Ching} and Lin, {Shun Chieh}",

year = "2009",

doi = "10.1109/CSIE.2009.739",

language = "???core.languages.en_GB???",

isbn = "9780769535074",

series = "2009 WRI World Congress on Computer Science and Information Engineering, CSIE 2009",

pages = "346--350",

booktitle = "2009 WRI World Congress on Computer Science and Information Engineering, CSIE 2009",

note = "2009 WRI World Congress on Computer Science and Information Engineering, CSIE 2009 ; Conference date: 31-03-2009 Through 02-04-2009",

}

Wang, JF, Chang, GM, Wang, JC & Lin, SC 2009, Stress detection based on multi-class probabilistic support vector machines for accented English speech. in 2009 WRI World Congress on Computer Science and Information Engineering, CSIE 2009., 5170340, 2009 WRI World Congress on Computer Science and Information Engineering, CSIE 2009, vol. 7, pp. 346-350, 2009 WRI World Congress on Computer Science and Information Engineering, CSIE 2009, Los Angeles, CA, United States, 31/03/09. https://doi.org/10.1109/CSIE.2009.739

Stress detection based on multi-class probabilistic support vector machines for accented English speech. / Wang, Jhing Fa; Chang, Gung Ming; Wang, Jia Ching et al.
2009 WRI World Congress on Computer Science and Information Engineering, CSIE 2009. 2009. p. 346-350 5170340 (2009 WRI World Congress on Computer Science and Information Engineering, CSIE 2009; Vol. 7).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Stress detection based on multi-class probabilistic support vector machines for accented English speech

AU - Wang, Jhing Fa

AU - Chang, Gung Ming

AU - Wang, Jia Ching

AU - Lin, Shun Chieh

PY - 2009

Y1 - 2009

N2 - A stress detection based on multi-class probabilistic support vector machines (MCP-SVMs) is proposed for classifying speech into following categories - no stress, primary stress, and secondary stress. The stress classifier is performed with a feature set including perceptual features, MFCC, delta-MFCC and delta-delta-MFCC. To observe that speakers from the same accent regions had similar tendencies in mispronunciations including word stress, this work uses English Across Taiwan (EAT) to represent Taiwanese-accented English speech corpora. The overall performance in the experimental results achieves about 84% classification of accuracy.

AB - A stress detection based on multi-class probabilistic support vector machines (MCP-SVMs) is proposed for classifying speech into following categories - no stress, primary stress, and secondary stress. The stress classifier is performed with a feature set including perceptual features, MFCC, delta-MFCC and delta-delta-MFCC. To observe that speakers from the same accent regions had similar tendencies in mispronunciations including word stress, this work uses English Across Taiwan (EAT) to represent Taiwanese-accented English speech corpora. The overall performance in the experimental results achieves about 84% classification of accuracy.

UR - http://www.scopus.com/inward/record.url?scp=71049120668&partnerID=8YFLogxK

U2 - 10.1109/CSIE.2009.739

DO - 10.1109/CSIE.2009.739

M3 - 會議論文篇章

AN - SCOPUS:71049120668

SN - 9780769535074

T3 - 2009 WRI World Congress on Computer Science and Information Engineering, CSIE 2009

SP - 346

EP - 350

BT - 2009 WRI World Congress on Computer Science and Information Engineering, CSIE 2009

T2 - 2009 WRI World Congress on Computer Science and Information Engineering, CSIE 2009

Y2 - 31 March 2009 through 2 April 2009

ER -

Stress detection based on multi-class probabilistic support vector machines for accented English speech

Abstract

Publication series

Conference

Access to Document

Fingerprint

Cite this