Skip to main content

Lund University Publications

LUND UNIVERSITY LIBRARIES

Offline Noise Reduction Using Optimal Mass Transport Induced Covariance Interpolation

Elvander, Filip LU ; Ali, Randall ; Jakobsson, Andreas LU orcid and van Waterschoot, Toon (2019) 27th European Signal Processing Conference (EUSIPCO)
Abstract
In this work, we propose to utilize a recently developed covariance matrix interpolation technique in order to improve noise reduction in multi-microphone setups in the presence of a moving, localized noise source. Based on the concept of optimal mass transport, the proposed method induces matrix interpolants implying smooth spatial displacement of the noise source, allowing for physically reasonable reconstructions of the noise source trajectory. As this trajectory is constructed as to connect two observed, or estimated, covariance matrices, the proposed method is suggested for offline applications. The performance of the proposed method is demonstrated using simulations of a speech enhancement scenario.
Please use this url to cite or link to this publication:
author
; ; and
organization
publishing date
type
Chapter in Book/Report/Conference proceeding
publication status
published
subject
host publication
27th European Signal Processing Conference, EUSIPCO 2019
pages
5 pages
publisher
IEEE - Institute of Electrical and Electronics Engineers Inc.
conference name
27th European Signal Processing Conference (EUSIPCO)
conference location
A Coruna, Spain
conference dates
2019-09-02 - 2019-09-06
external identifiers
  • scopus:85075626967
ISBN
978-1-5386-7300-3
978-9-0827-9703-9
DOI
10.23919/EUSIPCO.2019.8903159
language
English
LU publication?
yes
id
99df7682-3ad1-45c4-aa8d-cf1dc1c9b9f7
date added to LUP
2019-06-10 09:22:03
date last changed
2025-01-09 13:33:48
@inproceedings{99df7682-3ad1-45c4-aa8d-cf1dc1c9b9f7,
  abstract     = {{In this work, we propose to utilize a recently developed covariance matrix interpolation technique in order to improve noise reduction in multi-microphone setups in the presence of a moving, localized noise source. Based on the concept of optimal mass transport, the proposed method induces matrix interpolants implying smooth spatial displacement of the noise source, allowing for physically reasonable reconstructions of the noise source trajectory. As this trajectory is constructed as to connect two observed, or estimated, covariance matrices, the proposed method is suggested for offline applications. The performance of the proposed method is demonstrated using simulations of a speech enhancement scenario.}},
  author       = {{Elvander, Filip and Ali, Randall and Jakobsson, Andreas and van Waterschoot, Toon}},
  booktitle    = {{27th European Signal Processing Conference, EUSIPCO 2019}},
  isbn         = {{978-1-5386-7300-3}},
  language     = {{eng}},
  month        = {{11}},
  publisher    = {{IEEE - Institute of Electrical and Electronics Engineers Inc.}},
  title        = {{Offline Noise Reduction Using Optimal Mass Transport Induced Covariance Interpolation}},
  url          = {{http://dx.doi.org/10.23919/EUSIPCO.2019.8903159}},
  doi          = {{10.23919/EUSIPCO.2019.8903159}},
  year         = {{2019}},
}