Sorry, you need to enable JavaScript to visit this website.

facebooktwittermailshare

Deep Unfolding for Multichannel Source Separation

Title slide for Deep Unfolding for Multichannel Source Separation
Abstract: 

Deep unfolding has recently been proposed to derive novel deep network architectures from model-based approaches. In this paper, we consider its application to multichannel source separation. We unfold a multichannel Gaussian mixture model (MCGMM), resulting in a deep MCGMM computational network that directly processes complex-valued frequency-domain multichannel audio and has an architecture defined explicitly by a generative model, thus combining the advantages of deep networks and model-based approaches. We further extend the deep MCGMM by modeling the GMM states using an MRF, whose unfolded mean-field inference updates add dynamics across layers. Experiments on source separation for multichannel mixtures of two simultaneous speakers shows that the deep MCGMM leads to improved performance with respect to the original MCGMM model.

up
0 users have voted:

Paper Details

Authors:
Scott Wisdom, John R. Hershey, Jonathan Le Roux, Shinji Watanabe
Submitted On:
24 March 2016 - 9:11pm
Short Link:
Type:
Presentation Slides
Event:
Presenter's Name:
Scott Wisdom
Paper Code:
AASP-L5.01
Document Year:
2016
Cite

Document Files

WisdomHersheyLeRouxWatanabe_ICASSP2016_publish.pdf

(381 downloads)

Subscribe

[1] Scott Wisdom, John R. Hershey, Jonathan Le Roux, Shinji Watanabe, "Deep Unfolding for Multichannel Source Separation", IEEE SigPort, 2016. [Online]. Available: http://sigport.org/1031. Accessed: Nov. 16, 2018.
@article{1031-16,
url = {http://sigport.org/1031},
author = {Scott Wisdom; John R. Hershey; Jonathan Le Roux; Shinji Watanabe },
publisher = {IEEE SigPort},
title = {Deep Unfolding for Multichannel Source Separation},
year = {2016} }
TY - EJOUR
T1 - Deep Unfolding for Multichannel Source Separation
AU - Scott Wisdom; John R. Hershey; Jonathan Le Roux; Shinji Watanabe
PY - 2016
PB - IEEE SigPort
UR - http://sigport.org/1031
ER -
Scott Wisdom, John R. Hershey, Jonathan Le Roux, Shinji Watanabe. (2016). Deep Unfolding for Multichannel Source Separation. IEEE SigPort. http://sigport.org/1031
Scott Wisdom, John R. Hershey, Jonathan Le Roux, Shinji Watanabe, 2016. Deep Unfolding for Multichannel Source Separation. Available at: http://sigport.org/1031.
Scott Wisdom, John R. Hershey, Jonathan Le Roux, Shinji Watanabe. (2016). "Deep Unfolding for Multichannel Source Separation." Web.
1. Scott Wisdom, John R. Hershey, Jonathan Le Roux, Shinji Watanabe. Deep Unfolding for Multichannel Source Separation [Internet]. IEEE SigPort; 2016. Available from : http://sigport.org/1031