A Hierarchical Neural Network Model Based On A C/V Segmentation Algorithm For Isolated Mandarin Speech Recognition

Jhing Fa Wang; Chung Hsien Wu; Shih Hung Chang; Jau Yien Lee

doi:10.1109/78.134458

A Hierarchical Neural Network Model Based On A C/V Segmentation Algorithm For Isolated Mandarin Speech Recognition

Jhing Fa Wang, Chung Hsien Wu, Shih Hung Chang, Jau Yien Lee

Department of Computer Science and Information Engineering

Research output: Contribution to journal › Article › peer-review

60 Citations (Scopus)

Abstract

A novel algorithm simultaneously performing consonant/vowel (C/V) segmentation and pitch detection is proposed. Based on this algorithm, a consonant enhancement method and a hierarchical neural network scheme are explored for Mandarin speech recognition. As a result, an improvement of 12% in consonant recognition rate is obtained and the number of recognition candidates is reduced from 1300 to 63. A series of experiments over all Mandarin syllables (about 1300) are demonstrated in the speaker-dependent mode. Comparisons with the DTW algorithm are evaluated to show that the performance is satisfactory. An overall recognition rate of 90.14% is obtained.

Original language	English
Pages (from-to)	2141-2146
Number of pages	6
Journal	IEEE Transactions on Signal Processing
Volume	39
Issue number	9
DOIs	https://doi.org/10.1109/78.134458
Publication status	Published - 1991 Sept

All Science Journal Classification (ASJC) codes

Signal Processing
Electrical and Electronic Engineering

Access to Document

10.1109/78.134458

Cite this

@article{f042c3ff55d343c1b75f5fa4edf17319,

title = "A Hierarchical Neural Network Model Based On A C/V Segmentation Algorithm For Isolated Mandarin Speech Recognition",

abstract = "A novel algorithm simultaneously performing consonant/vowel (C/V) segmentation and pitch detection is proposed. Based on this algorithm, a consonant enhancement method and a hierarchical neural network scheme are explored for Mandarin speech recognition. As a result, an improvement of 12% in consonant recognition rate is obtained and the number of recognition candidates is reduced from 1300 to 63. A series of experiments over all Mandarin syllables (about 1300) are demonstrated in the speaker-dependent mode. Comparisons with the DTW algorithm are evaluated to show that the performance is satisfactory. An overall recognition rate of 90.14% is obtained.",

author = "Wang, {Jhing Fa} and Wu, {Chung Hsien} and Chang, {Shih Hung} and Lee, {Jau Yien}",

year = "1991",

month = sep,

doi = "10.1109/78.134458",

language = "English",

volume = "39",

pages = "2141--2146",

journal = "IEEE Transactions on Acoustics, Speech, and Signal Processing",

issn = "1053-587X",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

number = "9",

}

TY - JOUR

T1 - A Hierarchical Neural Network Model Based On A C/V Segmentation Algorithm For Isolated Mandarin Speech Recognition

AU - Wang, Jhing Fa

AU - Wu, Chung Hsien

AU - Chang, Shih Hung

AU - Lee, Jau Yien

PY - 1991/9

Y1 - 1991/9

N2 - A novel algorithm simultaneously performing consonant/vowel (C/V) segmentation and pitch detection is proposed. Based on this algorithm, a consonant enhancement method and a hierarchical neural network scheme are explored for Mandarin speech recognition. As a result, an improvement of 12% in consonant recognition rate is obtained and the number of recognition candidates is reduced from 1300 to 63. A series of experiments over all Mandarin syllables (about 1300) are demonstrated in the speaker-dependent mode. Comparisons with the DTW algorithm are evaluated to show that the performance is satisfactory. An overall recognition rate of 90.14% is obtained.

AB - A novel algorithm simultaneously performing consonant/vowel (C/V) segmentation and pitch detection is proposed. Based on this algorithm, a consonant enhancement method and a hierarchical neural network scheme are explored for Mandarin speech recognition. As a result, an improvement of 12% in consonant recognition rate is obtained and the number of recognition candidates is reduced from 1300 to 63. A series of experiments over all Mandarin syllables (about 1300) are demonstrated in the speaker-dependent mode. Comparisons with the DTW algorithm are evaluated to show that the performance is satisfactory. An overall recognition rate of 90.14% is obtained.

UR - http://www.scopus.com/inward/record.url?scp=0026219225&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0026219225&partnerID=8YFLogxK

U2 - 10.1109/78.134458

DO - 10.1109/78.134458

M3 - Article

AN - SCOPUS:0026219225

SN - 1053-587X

VL - 39

SP - 2141

EP - 2146

JO - IEEE Transactions on Acoustics, Speech, and Signal Processing

JF - IEEE Transactions on Acoustics, Speech, and Signal Processing

IS - 9

ER -

A Hierarchical Neural Network Model Based On A C/V Segmentation Algorithm For Isolated Mandarin Speech Recognition

Abstract

All Science Journal Classification (ASJC) codes

Access to Document

Other files and links

Fingerprint

Cite this