Speech enhancement using PCA and variance of the reconstruction error model identification

Amin Haji Abolhassani, Sid Ahmed Selouani, Douglas O'Shaughnessy, Mohamed-Faouzi Harkat

Research output: Chapter in Book/Report/Conference proceedingConference contribution

4 Citations (Scopus)

Abstract

We present in this paper a subspace approach for enhancing a noisy speech signal. The original algorithm for model identification from which we have derived our method has been used in the field of fault detection and diagnosis. This algorithm is based on principal component analysis in which the optimal subspace selection is provided by a variance of the reconstruction error (VRE) criterion. This choice overcomes many limitations encountered with other selection criteria, like overestimation of the signal subspace or the need for empirical parameters. We have also extended our subspace algorithm to take into account the case of colored and babble noise. The performance evaluation, which is made on the Aurora database shows that our method provides a higher noise reduction and a lower signal distortion than existing enhancement methods. Our algorithm succeeds in enhancing the noisy speech in all noisy conditions without introducing artifacts such as "musical noise".

Original languageEnglish
Title of host publicationInternational Speech Communication Association - 8th Annual Conference of the International Speech Communication Association, Interspeech 2007
Pages2384-2387
Number of pages4
Volume4
Publication statusPublished - 1 Dec 2007
Externally publishedYes
Event8th Annual Conference of the International Speech Communication Association, Interspeech 2007 - Antwerp, Belgium
Duration: 27 Aug 200731 Aug 2007

Other

Other8th Annual Conference of the International Speech Communication Association, Interspeech 2007
CountryBelgium
CityAntwerp
Period27/8/0731/8/07

Fingerprint

Speech Enhancement
Speech enhancement
Error Model
Model Identification
Identification (control systems)
reconstruction
Subspace
Fault Detection and Diagnosis
Signal distortion
Speech Signal
Noise Reduction
Noise abatement
Fault detection
Principal component analysis
Principal Component Analysis
Failure analysis
Performance Evaluation
artifact
Enhancement
evaluation

Keywords

  • Colored noise
  • Model identification
  • Principal component analysis
  • Signal subspace
  • Speech enhancement

ASJC Scopus subject areas

  • Computer Science Applications
  • Software
  • Modelling and Simulation
  • Linguistics and Language
  • Communication

Cite this

Abolhassani, A. H., Selouani, S. A., O'Shaughnessy, D., & Harkat, M-F. (2007). Speech enhancement using PCA and variance of the reconstruction error model identification. In International Speech Communication Association - 8th Annual Conference of the International Speech Communication Association, Interspeech 2007 (Vol. 4, pp. 2384-2387)

Speech enhancement using PCA and variance of the reconstruction error model identification. / Abolhassani, Amin Haji; Selouani, Sid Ahmed; O'Shaughnessy, Douglas; Harkat, Mohamed-Faouzi.

International Speech Communication Association - 8th Annual Conference of the International Speech Communication Association, Interspeech 2007. Vol. 4 2007. p. 2384-2387.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abolhassani, AH, Selouani, SA, O'Shaughnessy, D & Harkat, M-F 2007, Speech enhancement using PCA and variance of the reconstruction error model identification. in International Speech Communication Association - 8th Annual Conference of the International Speech Communication Association, Interspeech 2007. vol. 4, pp. 2384-2387, 8th Annual Conference of the International Speech Communication Association, Interspeech 2007, Antwerp, Belgium, 27/8/07.
Abolhassani AH, Selouani SA, O'Shaughnessy D, Harkat M-F. Speech enhancement using PCA and variance of the reconstruction error model identification. In International Speech Communication Association - 8th Annual Conference of the International Speech Communication Association, Interspeech 2007. Vol. 4. 2007. p. 2384-2387
Abolhassani, Amin Haji ; Selouani, Sid Ahmed ; O'Shaughnessy, Douglas ; Harkat, Mohamed-Faouzi. / Speech enhancement using PCA and variance of the reconstruction error model identification. International Speech Communication Association - 8th Annual Conference of the International Speech Communication Association, Interspeech 2007. Vol. 4 2007. pp. 2384-2387
@inproceedings{07e864905bef47a995f95cb8ad036f1a,
title = "Speech enhancement using PCA and variance of the reconstruction error model identification",
abstract = "We present in this paper a subspace approach for enhancing a noisy speech signal. The original algorithm for model identification from which we have derived our method has been used in the field of fault detection and diagnosis. This algorithm is based on principal component analysis in which the optimal subspace selection is provided by a variance of the reconstruction error (VRE) criterion. This choice overcomes many limitations encountered with other selection criteria, like overestimation of the signal subspace or the need for empirical parameters. We have also extended our subspace algorithm to take into account the case of colored and babble noise. The performance evaluation, which is made on the Aurora database shows that our method provides a higher noise reduction and a lower signal distortion than existing enhancement methods. Our algorithm succeeds in enhancing the noisy speech in all noisy conditions without introducing artifacts such as {"}musical noise{"}.",
keywords = "Colored noise, Model identification, Principal component analysis, Signal subspace, Speech enhancement",
author = "Abolhassani, {Amin Haji} and Selouani, {Sid Ahmed} and Douglas O'Shaughnessy and Mohamed-Faouzi Harkat",
year = "2007",
month = "12",
day = "1",
language = "English",
isbn = "9781605603162",
volume = "4",
pages = "2384--2387",
booktitle = "International Speech Communication Association - 8th Annual Conference of the International Speech Communication Association, Interspeech 2007",

}

TY - GEN

T1 - Speech enhancement using PCA and variance of the reconstruction error model identification

AU - Abolhassani, Amin Haji

AU - Selouani, Sid Ahmed

AU - O'Shaughnessy, Douglas

AU - Harkat, Mohamed-Faouzi

PY - 2007/12/1

Y1 - 2007/12/1

N2 - We present in this paper a subspace approach for enhancing a noisy speech signal. The original algorithm for model identification from which we have derived our method has been used in the field of fault detection and diagnosis. This algorithm is based on principal component analysis in which the optimal subspace selection is provided by a variance of the reconstruction error (VRE) criterion. This choice overcomes many limitations encountered with other selection criteria, like overestimation of the signal subspace or the need for empirical parameters. We have also extended our subspace algorithm to take into account the case of colored and babble noise. The performance evaluation, which is made on the Aurora database shows that our method provides a higher noise reduction and a lower signal distortion than existing enhancement methods. Our algorithm succeeds in enhancing the noisy speech in all noisy conditions without introducing artifacts such as "musical noise".

AB - We present in this paper a subspace approach for enhancing a noisy speech signal. The original algorithm for model identification from which we have derived our method has been used in the field of fault detection and diagnosis. This algorithm is based on principal component analysis in which the optimal subspace selection is provided by a variance of the reconstruction error (VRE) criterion. This choice overcomes many limitations encountered with other selection criteria, like overestimation of the signal subspace or the need for empirical parameters. We have also extended our subspace algorithm to take into account the case of colored and babble noise. The performance evaluation, which is made on the Aurora database shows that our method provides a higher noise reduction and a lower signal distortion than existing enhancement methods. Our algorithm succeeds in enhancing the noisy speech in all noisy conditions without introducing artifacts such as "musical noise".

KW - Colored noise

KW - Model identification

KW - Principal component analysis

KW - Signal subspace

KW - Speech enhancement

UR - http://www.scopus.com/inward/record.url?scp=56149118967&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=56149118967&partnerID=8YFLogxK

M3 - Conference contribution

SN - 9781605603162

VL - 4

SP - 2384

EP - 2387

BT - International Speech Communication Association - 8th Annual Conference of the International Speech Communication Association, Interspeech 2007

ER -