Forward-backward minima controlled recursive averaging to speech enhancement

Yeou Jiunn Chen; Jiunn Liang Wu

doi:10.1109/CIMSIVP.2013.6583847

Forward-backward minima controlled recursive averaging to speech enhancement

Yeou Jiunn Chen, Jiunn Liang Wu

耳鼻喉學科

研究成果: Conference contribution

2 引文斯高帕斯（Scopus）

摘要

The performance of hearing aids would be seriously reduced by nonstationary noises. In this paper, forward-backward minima controlled recursive averaging is proposed to improve the speech intelligibility by reducing the components of nonstationary noises. A forward-backward estimation is developed to correctively track nonstationary noises, especially the beginning of nonstationary noises. To suppress these nonstationary noises, a log-spectral amplitude estimator was selected as the gain function and used to effectively estimate the power spectrum of the noises. Moreover, harmonic regeneration was integrated to reduce the overestimation of noises. The experimental results demonstrated that our approach is able to improve performance compared to previous approaches.

原文	English
主出版物標題	Proceedings of the 2013 IEEE Symposium on Computational Intelligence for Multimedia, Signal and Vision Processing, CIMSIVP 2013 - 2013 IEEE Symposium Series on Computational Intelligence, SSCI 2013
頁面	49-52
頁數	4
DOIs	https://doi.org/10.1109/CIMSIVP.2013.6583847
出版狀態	Published - 2013
事件	2013 IEEE Symposium on Computational Intelligence for Multimedia, Signal and Vision Processing, CIMSIVP 2013 - 2013 IEEE Symposium Series on Computational Intelligence, SSCI 2013 - Singapore, Singapore 持續時間: 2013 4月 16 → 2013 4月 19

出版系列

名字	Proceedings of the 2013 IEEE Symposium on Computational Intelligence for Multimedia, Signal and Vision Processing, CIMSIVP 2013 - 2013 IEEE Symposium Series on Computational Intelligence, SSCI 2013

Other

Other	2013 IEEE Symposium on Computational Intelligence for Multimedia, Signal and Vision Processing, CIMSIVP 2013 - 2013 IEEE Symposium Series on Computational Intelligence, SSCI 2013
國家/地區	Singapore
城市	Singapore
期間	13-04-16 → 13-04-19

All Science Journal Classification (ASJC) codes

電腦繪圖與電腦輔助設計
電腦視覺和模式識別

UN SDG

此研究成果有助於以下永續發展目標

存取文件

10.1109/CIMSIVP.2013.6583847

其它文件與鏈接

引用此

Chen, Y. J., & Wu, J. L. (2013). Forward-backward minima controlled recursive averaging to speech enhancement. 於 Proceedings of the 2013 IEEE Symposium on Computational Intelligence for Multimedia, Signal and Vision Processing, CIMSIVP 2013 - 2013 IEEE Symposium Series on Computational Intelligence, SSCI 2013 (頁 49-52). 文章 6583847 (Proceedings of the 2013 IEEE Symposium on Computational Intelligence for Multimedia, Signal and Vision Processing, CIMSIVP 2013 - 2013 IEEE Symposium Series on Computational Intelligence, SSCI 2013). https://doi.org/10.1109/CIMSIVP.2013.6583847

Chen, Yeou Jiunn ; Wu, Jiunn Liang. / Forward-backward minima controlled recursive averaging to speech enhancement. Proceedings of the 2013 IEEE Symposium on Computational Intelligence for Multimedia, Signal and Vision Processing, CIMSIVP 2013 - 2013 IEEE Symposium Series on Computational Intelligence, SSCI 2013. 2013. 頁 49-52 (Proceedings of the 2013 IEEE Symposium on Computational Intelligence for Multimedia, Signal and Vision Processing, CIMSIVP 2013 - 2013 IEEE Symposium Series on Computational Intelligence, SSCI 2013).

@inproceedings{57b86f0f9f2a494fbab59f36b268a12b,

title = "Forward-backward minima controlled recursive averaging to speech enhancement",

abstract = "The performance of hearing aids would be seriously reduced by nonstationary noises. In this paper, forward-backward minima controlled recursive averaging is proposed to improve the speech intelligibility by reducing the components of nonstationary noises. A forward-backward estimation is developed to correctively track nonstationary noises, especially the beginning of nonstationary noises. To suppress these nonstationary noises, a log-spectral amplitude estimator was selected as the gain function and used to effectively estimate the power spectrum of the noises. Moreover, harmonic regeneration was integrated to reduce the overestimation of noises. The experimental results demonstrated that our approach is able to improve performance compared to previous approaches.",

author = "Chen, {Yeou Jiunn} and Wu, {Jiunn Liang}",

year = "2013",

doi = "10.1109/CIMSIVP.2013.6583847",

language = "English",

isbn = "9781467359177",

series = "Proceedings of the 2013 IEEE Symposium on Computational Intelligence for Multimedia, Signal and Vision Processing, CIMSIVP 2013 - 2013 IEEE Symposium Series on Computational Intelligence, SSCI 2013",

pages = "49--52",

booktitle = "Proceedings of the 2013 IEEE Symposium on Computational Intelligence for Multimedia, Signal and Vision Processing, CIMSIVP 2013 - 2013 IEEE Symposium Series on Computational Intelligence, SSCI 2013",

note = "2013 IEEE Symposium on Computational Intelligence for Multimedia, Signal and Vision Processing, CIMSIVP 2013 - 2013 IEEE Symposium Series on Computational Intelligence, SSCI 2013 ; Conference date: 16-04-2013 Through 19-04-2013",

}

Chen, YJ & Wu, JL 2013, Forward-backward minima controlled recursive averaging to speech enhancement. 於 Proceedings of the 2013 IEEE Symposium on Computational Intelligence for Multimedia, Signal and Vision Processing, CIMSIVP 2013 - 2013 IEEE Symposium Series on Computational Intelligence, SSCI 2013., 6583847, Proceedings of the 2013 IEEE Symposium on Computational Intelligence for Multimedia, Signal and Vision Processing, CIMSIVP 2013 - 2013 IEEE Symposium Series on Computational Intelligence, SSCI 2013, 頁 49-52, 2013 IEEE Symposium on Computational Intelligence for Multimedia, Signal and Vision Processing, CIMSIVP 2013 - 2013 IEEE Symposium Series on Computational Intelligence, SSCI 2013, Singapore, Singapore, 13-04-16. https://doi.org/10.1109/CIMSIVP.2013.6583847

Forward-backward minima controlled recursive averaging to speech enhancement. / Chen, Yeou Jiunn; Wu, Jiunn Liang.
Proceedings of the 2013 IEEE Symposium on Computational Intelligence for Multimedia, Signal and Vision Processing, CIMSIVP 2013 - 2013 IEEE Symposium Series on Computational Intelligence, SSCI 2013. 2013. p. 49-52 6583847 (Proceedings of the 2013 IEEE Symposium on Computational Intelligence for Multimedia, Signal and Vision Processing, CIMSIVP 2013 - 2013 IEEE Symposium Series on Computational Intelligence, SSCI 2013).

研究成果: Conference contribution

TY - GEN

T1 - Forward-backward minima controlled recursive averaging to speech enhancement

AU - Chen, Yeou Jiunn

AU - Wu, Jiunn Liang

PY - 2013

Y1 - 2013

N2 - The performance of hearing aids would be seriously reduced by nonstationary noises. In this paper, forward-backward minima controlled recursive averaging is proposed to improve the speech intelligibility by reducing the components of nonstationary noises. A forward-backward estimation is developed to correctively track nonstationary noises, especially the beginning of nonstationary noises. To suppress these nonstationary noises, a log-spectral amplitude estimator was selected as the gain function and used to effectively estimate the power spectrum of the noises. Moreover, harmonic regeneration was integrated to reduce the overestimation of noises. The experimental results demonstrated that our approach is able to improve performance compared to previous approaches.

AB - The performance of hearing aids would be seriously reduced by nonstationary noises. In this paper, forward-backward minima controlled recursive averaging is proposed to improve the speech intelligibility by reducing the components of nonstationary noises. A forward-backward estimation is developed to correctively track nonstationary noises, especially the beginning of nonstationary noises. To suppress these nonstationary noises, a log-spectral amplitude estimator was selected as the gain function and used to effectively estimate the power spectrum of the noises. Moreover, harmonic regeneration was integrated to reduce the overestimation of noises. The experimental results demonstrated that our approach is able to improve performance compared to previous approaches.

UR - http://www.scopus.com/inward/record.url?scp=84883685056&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84883685056&partnerID=8YFLogxK

U2 - 10.1109/CIMSIVP.2013.6583847

DO - 10.1109/CIMSIVP.2013.6583847

M3 - Conference contribution

AN - SCOPUS:84883685056

SN - 9781467359177

T3 - Proceedings of the 2013 IEEE Symposium on Computational Intelligence for Multimedia, Signal and Vision Processing, CIMSIVP 2013 - 2013 IEEE Symposium Series on Computational Intelligence, SSCI 2013

SP - 49

EP - 52

BT - Proceedings of the 2013 IEEE Symposium on Computational Intelligence for Multimedia, Signal and Vision Processing, CIMSIVP 2013 - 2013 IEEE Symposium Series on Computational Intelligence, SSCI 2013

T2 - 2013 IEEE Symposium on Computational Intelligence for Multimedia, Signal and Vision Processing, CIMSIVP 2013 - 2013 IEEE Symposium Series on Computational Intelligence, SSCI 2013

Y2 - 16 April 2013 through 19 April 2013

ER -

Chen YJ, Wu JL. Forward-backward minima controlled recursive averaging to speech enhancement. 於 Proceedings of the 2013 IEEE Symposium on Computational Intelligence for Multimedia, Signal and Vision Processing, CIMSIVP 2013 - 2013 IEEE Symposium Series on Computational Intelligence, SSCI 2013. 2013. p. 49-52. 6583847. (Proceedings of the 2013 IEEE Symposium on Computational Intelligence for Multimedia, Signal and Vision Processing, CIMSIVP 2013 - 2013 IEEE Symposium Series on Computational Intelligence, SSCI 2013). doi: 10.1109/CIMSIVP.2013.6583847

Forward-backward minima controlled recursive averaging to speech enhancement

摘要

出版系列

Other

All Science Journal Classification (ASJC) codes

UN SDG

存取文件

其它文件與鏈接

指紋

引用此