Offline Noise Reduction Using Optimal Mass Transport Induced Covariance Interpolation

Filip Elvander; Randall Ali; Andreas Jakobsson; Toon van Waterschoot

doi:10.23919/EUSIPCO.2019.8903159

Offline Noise Reduction Using Optimal Mass Transport Induced Covariance Interpolation

Filip Elvander, Randall Ali, Andreas Jakobsson, Toon van Waterschoot

Catholic University of Leuven

Research output: Chapter in Book/Report/Conference proceeding › Paper in conference proceeding › peer-review

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.

Original language	English
Title of host publication	27th European Signal Processing Conference, EUSIPCO 2019
Publisher	IEEE - Institute of Electrical and Electronics Engineers Inc.
Number of pages	5
ISBN (Electronic)	978-9-0827-9703-9
ISBN (Print)	978-1-5386-7300-3
DOIs	https://doi.org/10.23919/EUSIPCO.2019.8903159
Publication status	Published - 2019 Nov 18
Event	27th European Signal Processing Conference (EUSIPCO) - A Coruna, Spain Duration: 2019 Sept 2 → 2019 Sept 6

Conference

Conference	27th European Signal Processing Conference (EUSIPCO)
Country/Territory	Spain
City	A Coruna
Period	2019/09/02 → 2019/09/06

Subject classification (UKÄ)

Signal Processing

Access to Document

10.23919/EUSIPCO.2019.8903159

Cite this

@inproceedings{99df76823ad145c4aa8dcf1dc1c9b9f7,

title = "Offline Noise Reduction Using Optimal Mass Transport Induced Covariance Interpolation",

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 = "Filip Elvander and Randall Ali and Andreas Jakobsson and {van Waterschoot}, Toon",

year = "2019",

month = nov,

day = "18",

doi = "10.23919/EUSIPCO.2019.8903159",

language = "English",

isbn = "978-1-5386-7300-3",

booktitle = "27th European Signal Processing Conference, EUSIPCO 2019",

publisher = "IEEE - Institute of Electrical and Electronics Engineers Inc.",

address = "United States",

note = "27th European Signal Processing Conference (EUSIPCO) ; Conference date: 02-09-2019 Through 06-09-2019",

}

Elvander, F, Ali, R, Jakobsson, A & van Waterschoot, T 2019, Offline Noise Reduction Using Optimal Mass Transport Induced Covariance Interpolation. in 27th European Signal Processing Conference, EUSIPCO 2019. IEEE - Institute of Electrical and Electronics Engineers Inc., 27th European Signal Processing Conference (EUSIPCO), A Coruna, Spain, 2019/09/02. https://doi.org/10.23919/EUSIPCO.2019.8903159

Offline Noise Reduction Using Optimal Mass Transport Induced Covariance Interpolation. / Elvander, Filip; Ali, Randall; Jakobsson, Andreas et al.

27th European Signal Processing Conference, EUSIPCO 2019. IEEE - Institute of Electrical and Electronics Engineers Inc., 2019.

Research output: Chapter in Book/Report/Conference proceeding › Paper in conference proceeding › peer-review

TY - GEN

T1 - Offline Noise Reduction Using Optimal Mass Transport Induced Covariance Interpolation

AU - Elvander, Filip

AU - Ali, Randall

AU - Jakobsson, Andreas

AU - van Waterschoot, Toon

PY - 2019/11/18

Y1 - 2019/11/18

N2 - 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.

AB - 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.

UR - https://scopus.com/record/display.uri?eid=2-s2.0-85075626967&origin=inward&txGid

U2 - 10.23919/EUSIPCO.2019.8903159

DO - 10.23919/EUSIPCO.2019.8903159

M3 - Paper in conference proceeding

SN - 978-1-5386-7300-3

BT - 27th European Signal Processing Conference, EUSIPCO 2019

PB - IEEE - Institute of Electrical and Electronics Engineers Inc.

T2 - 27th European Signal Processing Conference (EUSIPCO)

Y2 - 2 September 2019 through 6 September 2019

ER -

Offline Noise Reduction Using Optimal Mass Transport Induced Covariance Interpolation

Abstract

Conference

Subject classification (UKÄ)

Access to Document

Other files and links

Fingerprint

Cite this