Time-frequency clustering with weighted and contextual information for convolutive blind source separation

I. Jafari, M. Atcheson, Roberto Togneri, S.E. Nordholm

Research output: Chapter in Book/Conference paperConference paper

Abstract

In this paper we investigate the use of observation weights and contextual time-frequency information for clustering-based blind source separation. Previous clustering-based approaches have successfully used clustering techniques to estimate time-frequency separation masks; however, these approaches generally disregard the structured nature of speech signals. Motivated by the homogenous behavior of speech signals, we propose to modify the established fuzzy c-means algorithm to bias the clustering results in favor of cluster membership homogeneity within localized neighborhoods in the time-frequency space. This problem can be solved by using a two-stage algorithm: firstly, the estimation of data weights to indicate the reliability of each data point, and secondly, the integration of local contextual information into the cluster update equations from neighboring time-frequency slots. The proposed algorithm is evaluated in a three-fold manner using simulated, real recordings and public benchmark data; notable improvement in source separation performance over previous clustering approaches was achieved. © 2014 IEEE.
Original languageEnglish
Title of host publicationIEEE International Conference on Acoustics, Speech and Signal Processing
Place of PublicationFlorence, Italy
PublisherIEEE, Institute of Electrical and Electronics Engineers
Pages157-160
ISBN (Print)9781479949755
DOIs
Publication statusPublished - 2014
Event2014 IEEE Workshop on Statistical Signal Processing - Gold Coast, Australia
Duration: 29 Jun 20142 Jul 2014

Conference

Conference2014 IEEE Workshop on Statistical Signal Processing
CountryAustralia
CityGold Coast
Period29/06/142/07/14

Fingerprint

Blind source separation
Source separation
Masks

Cite this

Jafari, I., Atcheson, M., Togneri, R., & Nordholm, S. E. (2014). Time-frequency clustering with weighted and contextual information for convolutive blind source separation. In IEEE International Conference on Acoustics, Speech and Signal Processing (pp. 157-160). Florence, Italy: IEEE, Institute of Electrical and Electronics Engineers. https://doi.org/10.1109/SSP.2014.6884599
Jafari, I. ; Atcheson, M. ; Togneri, Roberto ; Nordholm, S.E. / Time-frequency clustering with weighted and contextual information for convolutive blind source separation. IEEE International Conference on Acoustics, Speech and Signal Processing. Florence, Italy : IEEE, Institute of Electrical and Electronics Engineers, 2014. pp. 157-160
@inproceedings{0a500373afa5491c98f423f07aa354c0,
title = "Time-frequency clustering with weighted and contextual information for convolutive blind source separation",
abstract = "In this paper we investigate the use of observation weights and contextual time-frequency information for clustering-based blind source separation. Previous clustering-based approaches have successfully used clustering techniques to estimate time-frequency separation masks; however, these approaches generally disregard the structured nature of speech signals. Motivated by the homogenous behavior of speech signals, we propose to modify the established fuzzy c-means algorithm to bias the clustering results in favor of cluster membership homogeneity within localized neighborhoods in the time-frequency space. This problem can be solved by using a two-stage algorithm: firstly, the estimation of data weights to indicate the reliability of each data point, and secondly, the integration of local contextual information into the cluster update equations from neighboring time-frequency slots. The proposed algorithm is evaluated in a three-fold manner using simulated, real recordings and public benchmark data; notable improvement in source separation performance over previous clustering approaches was achieved. {\circledC} 2014 IEEE.",
author = "I. Jafari and M. Atcheson and Roberto Togneri and S.E. Nordholm",
year = "2014",
doi = "10.1109/SSP.2014.6884599",
language = "English",
isbn = "9781479949755",
pages = "157--160",
booktitle = "IEEE International Conference on Acoustics, Speech and Signal Processing",
publisher = "IEEE, Institute of Electrical and Electronics Engineers",
address = "United States",

}

Jafari, I, Atcheson, M, Togneri, R & Nordholm, SE 2014, Time-frequency clustering with weighted and contextual information for convolutive blind source separation. in IEEE International Conference on Acoustics, Speech and Signal Processing. IEEE, Institute of Electrical and Electronics Engineers, Florence, Italy, pp. 157-160, 2014 IEEE Workshop on Statistical Signal Processing, Gold Coast, Australia, 29/06/14. https://doi.org/10.1109/SSP.2014.6884599

Time-frequency clustering with weighted and contextual information for convolutive blind source separation. / Jafari, I.; Atcheson, M.; Togneri, Roberto; Nordholm, S.E.

IEEE International Conference on Acoustics, Speech and Signal Processing. Florence, Italy : IEEE, Institute of Electrical and Electronics Engineers, 2014. p. 157-160.

Research output: Chapter in Book/Conference paperConference paper

TY - GEN

T1 - Time-frequency clustering with weighted and contextual information for convolutive blind source separation

AU - Jafari, I.

AU - Atcheson, M.

AU - Togneri, Roberto

AU - Nordholm, S.E.

PY - 2014

Y1 - 2014

N2 - In this paper we investigate the use of observation weights and contextual time-frequency information for clustering-based blind source separation. Previous clustering-based approaches have successfully used clustering techniques to estimate time-frequency separation masks; however, these approaches generally disregard the structured nature of speech signals. Motivated by the homogenous behavior of speech signals, we propose to modify the established fuzzy c-means algorithm to bias the clustering results in favor of cluster membership homogeneity within localized neighborhoods in the time-frequency space. This problem can be solved by using a two-stage algorithm: firstly, the estimation of data weights to indicate the reliability of each data point, and secondly, the integration of local contextual information into the cluster update equations from neighboring time-frequency slots. The proposed algorithm is evaluated in a three-fold manner using simulated, real recordings and public benchmark data; notable improvement in source separation performance over previous clustering approaches was achieved. © 2014 IEEE.

AB - In this paper we investigate the use of observation weights and contextual time-frequency information for clustering-based blind source separation. Previous clustering-based approaches have successfully used clustering techniques to estimate time-frequency separation masks; however, these approaches generally disregard the structured nature of speech signals. Motivated by the homogenous behavior of speech signals, we propose to modify the established fuzzy c-means algorithm to bias the clustering results in favor of cluster membership homogeneity within localized neighborhoods in the time-frequency space. This problem can be solved by using a two-stage algorithm: firstly, the estimation of data weights to indicate the reliability of each data point, and secondly, the integration of local contextual information into the cluster update equations from neighboring time-frequency slots. The proposed algorithm is evaluated in a three-fold manner using simulated, real recordings and public benchmark data; notable improvement in source separation performance over previous clustering approaches was achieved. © 2014 IEEE.

U2 - 10.1109/SSP.2014.6884599

DO - 10.1109/SSP.2014.6884599

M3 - Conference paper

SN - 9781479949755

SP - 157

EP - 160

BT - IEEE International Conference on Acoustics, Speech and Signal Processing

PB - IEEE, Institute of Electrical and Electronics Engineers

CY - Florence, Italy

ER -

Jafari I, Atcheson M, Togneri R, Nordholm SE. Time-frequency clustering with weighted and contextual information for convolutive blind source separation. In IEEE International Conference on Acoustics, Speech and Signal Processing. Florence, Italy: IEEE, Institute of Electrical and Electronics Engineers. 2014. p. 157-160 https://doi.org/10.1109/SSP.2014.6884599