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

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 separationmasks; however, these approaches...

Full description

Bibliographic Details
Main Authors: Jafari, I., Atcheson, M., Togneri, R., Nordholm, Sven
Other Authors: Sergios Theodoridis
Format: Conference Paper
Published: Institute of Electrical and Electronics Engineers ( IEEE ) 2014
Subjects:
Online Access:http://hdl.handle.net/20.500.11937/28674
_version_ 1848752600125538304
author Jafari, I.
Atcheson, M.
Togneri, R.
Nordholm, Sven
author2 Sergios Theodoridis
author_facet Sergios Theodoridis
Jafari, I.
Atcheson, M.
Togneri, R.
Nordholm, Sven
author_sort Jafari, I.
building Curtin Institutional Repository
collection Online Access
description 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 separationmasks; however, these approaches generally disregard the structured nature of speech signals. Motivated by the homogenous behaviour of speech signals, we propose to modify the established fuzzy cmeans 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.
first_indexed 2025-11-14T08:11:12Z
format Conference Paper
id curtin-20.500.11937-28674
institution Curtin University Malaysia
institution_category Local University
last_indexed 2025-11-14T08:11:12Z
publishDate 2014
publisher Institute of Electrical and Electronics Engineers ( IEEE )
recordtype eprints
repository_type Digital Repository
spelling curtin-20.500.11937-286742017-09-13T15:15:33Z Time-frequency clustering with weighted and contextual information for convolutive blind source separation Jafari, I. Atcheson, M. Togneri, R. Nordholm, Sven Sergios Theodoridis fuzzy c-means clustering time-frequency masking observation weights blind source separation contextual information 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 separationmasks; however, these approaches generally disregard the structured nature of speech signals. Motivated by the homogenous behaviour of speech signals, we propose to modify the established fuzzy cmeans 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 Conference Paper http://hdl.handle.net/20.500.11937/28674 10.1109/SSP.2014.6884599 Institute of Electrical and Electronics Engineers ( IEEE ) restricted
spellingShingle fuzzy c-means clustering
time-frequency masking
observation weights
blind source separation
contextual information
Jafari, I.
Atcheson, M.
Togneri, R.
Nordholm, Sven
Time-frequency clustering with weighted and contextual information for convolutive blind source separation
title Time-frequency clustering with weighted and contextual information for convolutive blind source separation
title_full Time-frequency clustering with weighted and contextual information for convolutive blind source separation
title_fullStr Time-frequency clustering with weighted and contextual information for convolutive blind source separation
title_full_unstemmed Time-frequency clustering with weighted and contextual information for convolutive blind source separation
title_short Time-frequency clustering with weighted and contextual information for convolutive blind source separation
title_sort time-frequency clustering with weighted and contextual information for convolutive blind source separation
topic fuzzy c-means clustering
time-frequency masking
observation weights
blind source separation
contextual information
url http://hdl.handle.net/20.500.11937/28674