Real-time independent vector analysis using semi-supervised nonnegative matrix factorization as a source model

Taihui Wang; Feiran Yang; Rui Zhu; Jun Yang

doi:10.21437/Interspeech.2021-146

Real-time independent vector analysis using semi-supervised nonnegative matrix factorization as a source model

Taihui Wang, Feiran Yang, Rui Zhu, Jun Yang

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

Abstract

Online independent vector analysis (IVA) based on auxiliary technology is effective to separate audio source in real time. However, the separated signal may contain residual interference noise because the source model of IVA lacks flexibility and cannot treat the specific harmonic structures of sources. This paper presents a real-time IVA method where the amplitude spectrum of separated signal is modeled by semi-supervised nonnegative matrix factorization (SSNMF). Using the pre-trained basis matrix which contains source structures, we can extract the target source from the separated signal in real time. The advantage of the proposed method is that the extracted source can provide a more accurate variance than the separated signal and hence the proposed method can obtain a better separation performance than the oracle IVA. Experimental results in speech denoising task show the effectiveness and the robustness of the proposed method with different types of noise.

Original language	English
Title of host publication	22nd Annual Conference of the International Speech Communication Association, INTERSPEECH 2021
Publisher	International Speech Communication Association
Pages	336-340
Number of pages	5
ISBN (Electronic)	9781713836902
DOIs	https://doi.org/10.21437/Interspeech.2021-146
Publication status	Published - 2021
Externally published	Yes
Event	22nd Annual Conference of the International Speech Communication Association, INTERSPEECH 2021 - Brno, Czech Republic Duration: Aug 30 2021 → Sept 3 2021

Publication series

Name	Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH
Volume	1
ISSN (Print)	2308-457X
ISSN (Electronic)	1990-9772

Conference

Conference	22nd Annual Conference of the International Speech Communication Association, INTERSPEECH 2021
Country/Territory	Czech Republic
City	Brno
Period	8/30/21 → 9/3/21

Bibliographical note

Publisher Copyright:
Copyright © 2021 ISCA.

ASJC Scopus Subject Areas

Language and Linguistics
Human-Computer Interaction
Signal Processing
Software
Modelling and Simulation

Keywords

Audio source separation
Realtime independent vector analysis
Semi-supervised nonnegative matrix factorization
Speech denoising

Access to Document

10.21437/Interspeech.2021-146

Cite this

Wang, T., Yang, F., Zhu, R., & Yang, J. (2021). Real-time independent vector analysis using semi-supervised nonnegative matrix factorization as a source model. In 22nd Annual Conference of the International Speech Communication Association, INTERSPEECH 2021 (pp. 336-340). (Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH; Vol. 1). International Speech Communication Association. https://doi.org/10.21437/Interspeech.2021-146

Wang, Taihui ; Yang, Feiran ; Zhu, Rui et al. / Real-time independent vector analysis using semi-supervised nonnegative matrix factorization as a source model. 22nd Annual Conference of the International Speech Communication Association, INTERSPEECH 2021. International Speech Communication Association, 2021. pp. 336-340 (Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH).

@inproceedings{3c508efa4e1d49e980f518f4872c3bf2,

title = "Real-time independent vector analysis using semi-supervised nonnegative matrix factorization as a source model",

abstract = "Online independent vector analysis (IVA) based on auxiliary technology is effective to separate audio source in real time. However, the separated signal may contain residual interference noise because the source model of IVA lacks flexibility and cannot treat the specific harmonic structures of sources. This paper presents a real-time IVA method where the amplitude spectrum of separated signal is modeled by semi-supervised nonnegative matrix factorization (SSNMF). Using the pre-trained basis matrix which contains source structures, we can extract the target source from the separated signal in real time. The advantage of the proposed method is that the extracted source can provide a more accurate variance than the separated signal and hence the proposed method can obtain a better separation performance than the oracle IVA. Experimental results in speech denoising task show the effectiveness and the robustness of the proposed method with different types of noise.",

keywords = "Audio source separation, Realtime independent vector analysis, Semi-supervised nonnegative matrix factorization, Speech denoising",

author = "Taihui Wang and Feiran Yang and Rui Zhu and Jun Yang",

note = "Publisher Copyright: Copyright {\textcopyright} 2021 ISCA.; 22nd Annual Conference of the International Speech Communication Association, INTERSPEECH 2021 ; Conference date: 30-08-2021 Through 03-09-2021",

year = "2021",

doi = "10.21437/Interspeech.2021-146",

language = "English",

series = "Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH",

publisher = "International Speech Communication Association",

pages = "336--340",

booktitle = "22nd Annual Conference of the International Speech Communication Association, INTERSPEECH 2021",

}

Wang, T, Yang, F, Zhu, R & Yang, J 2021, Real-time independent vector analysis using semi-supervised nonnegative matrix factorization as a source model. in 22nd Annual Conference of the International Speech Communication Association, INTERSPEECH 2021. Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH, vol. 1, International Speech Communication Association, pp. 336-340, 22nd Annual Conference of the International Speech Communication Association, INTERSPEECH 2021, Brno, Czech Republic, 8/30/21. https://doi.org/10.21437/Interspeech.2021-146

Real-time independent vector analysis using semi-supervised nonnegative matrix factorization as a source model. / Wang, Taihui; Yang, Feiran; Zhu, Rui et al.
22nd Annual Conference of the International Speech Communication Association, INTERSPEECH 2021. International Speech Communication Association, 2021. p. 336-340 (Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH; Vol. 1).

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

TY - GEN

T1 - Real-time independent vector analysis using semi-supervised nonnegative matrix factorization as a source model

AU - Wang, Taihui

AU - Yang, Feiran

AU - Zhu, Rui

AU - Yang, Jun

PY - 2021

Y1 - 2021

N2 - Online independent vector analysis (IVA) based on auxiliary technology is effective to separate audio source in real time. However, the separated signal may contain residual interference noise because the source model of IVA lacks flexibility and cannot treat the specific harmonic structures of sources. This paper presents a real-time IVA method where the amplitude spectrum of separated signal is modeled by semi-supervised nonnegative matrix factorization (SSNMF). Using the pre-trained basis matrix which contains source structures, we can extract the target source from the separated signal in real time. The advantage of the proposed method is that the extracted source can provide a more accurate variance than the separated signal and hence the proposed method can obtain a better separation performance than the oracle IVA. Experimental results in speech denoising task show the effectiveness and the robustness of the proposed method with different types of noise.

AB - Online independent vector analysis (IVA) based on auxiliary technology is effective to separate audio source in real time. However, the separated signal may contain residual interference noise because the source model of IVA lacks flexibility and cannot treat the specific harmonic structures of sources. This paper presents a real-time IVA method where the amplitude spectrum of separated signal is modeled by semi-supervised nonnegative matrix factorization (SSNMF). Using the pre-trained basis matrix which contains source structures, we can extract the target source from the separated signal in real time. The advantage of the proposed method is that the extracted source can provide a more accurate variance than the separated signal and hence the proposed method can obtain a better separation performance than the oracle IVA. Experimental results in speech denoising task show the effectiveness and the robustness of the proposed method with different types of noise.

KW - Audio source separation

KW - Realtime independent vector analysis

KW - Semi-supervised nonnegative matrix factorization

KW - Speech denoising

UR - http://www.scopus.com/inward/record.url?scp=85119300071&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85119300071&partnerID=8YFLogxK

U2 - 10.21437/Interspeech.2021-146

DO - 10.21437/Interspeech.2021-146

M3 - Conference contribution

AN - SCOPUS:85119300071

T3 - Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH

SP - 336

EP - 340

BT - 22nd Annual Conference of the International Speech Communication Association, INTERSPEECH 2021

PB - International Speech Communication Association

T2 - 22nd Annual Conference of the International Speech Communication Association, INTERSPEECH 2021

Y2 - 30 August 2021 through 3 September 2021

ER -

Wang T, Yang F, Zhu R, Yang J. Real-time independent vector analysis using semi-supervised nonnegative matrix factorization as a source model. In 22nd Annual Conference of the International Speech Communication Association, INTERSPEECH 2021. International Speech Communication Association. 2021. p. 336-340. (Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH). doi: 10.21437/Interspeech.2021-146

Real-time independent vector analysis using semi-supervised nonnegative matrix factorization as a source model

Abstract

Publication series

Conference

Bibliographical note

ASJC Scopus Subject Areas

Keywords

Access to Document

Other files and links

Fingerprint

Cite this