Article (Scientific journals)
Multichannel Source Separation Using Time-Deconvolutive CNMF
Dias, Thadeu; Alves Martins, Wallace; Biscainho, Luiz Wagner
2020In Journal of Communication and Information Systems, 35 (1), p. 103-112
Peer reviewed
 

Files


Full Text
707-Article Text-3075-1-10-20200514.pdf
Publisher postprint (1.6 MB)
Download

All documents in ORBilu are protected by a user license.

Send to



Details



Keywords :
Blind source separation; convolutive mixture; NMF; deconvolutive NMF
Abstract :
[en] This paper addresses the separation of audio sources from convolutive mixtures captured by a microphone array. We approach the problem using complex-valued non-negative matrix factorization (CNMF), and extend previous works by tailoring advanced (single-channel) NMF models, such as the deconvolutive NMF, to the multichannel factorization setup. Further, a sparsity-promoting scheme is proposed so that the underlying estimated parameters better fit the time-frequency properties inherent in some audio sources. The proposed parameter estimation framework is compatible with previous related works, and can be thought of as a step toward a more general method. We evaluate the resulting separation accuracy using a simulated acoustic scenario, and the tests confirm that the proposed algorithm provides superior separation quality when compared to a state-of-the-art benchmark. Finally, an analysis of the effects of the introduced regularization term shows that the solution is in fact steered toward a sparser representation.
Research center :
Interdisciplinary Centre for Security, Reliability and Trust (SnT) > SIGCOM
Disciplines :
Electrical & electronics engineering
Author, co-author :
Dias, Thadeu;  Federal University of Rio de Janeiro (UFRJ)
Alves Martins, Wallace ;  University of Luxembourg > Interdisciplinary Centre for Security, Reliability and Trust (SNT)
Biscainho, Luiz Wagner;  Federal University of Rio de Janeiro (UFRJ)
External co-authors :
yes
Language :
English
Title :
Multichannel Source Separation Using Time-Deconvolutive CNMF
Publication date :
May 2020
Journal title :
Journal of Communication and Information Systems
ISSN :
1980-6604
Volume :
35
Issue :
1
Pages :
103-112
Peer reviewed :
Peer reviewed
Focus Area :
Security, Reliability and Trust
Available on ORBilu :
since 26 May 2020

Statistics


Number of views
53 (0 by Unilu)
Number of downloads
37 (0 by Unilu)

OpenCitations
 
2

Bibliography


Similar publications



Contact ORBilu