Exploiting structures of temporal causality for robust speaker localization in reverberant environments

Christopher Schymura; Peng Guo; Yanir Maymon; Boaz Rafaely; Dorothea Kolossa

doi:https://doi.org/10.1007/978-3-319-93764-9_22

Exploiting structures of temporal causality for robust speaker localization in reverberant environments

Christopher Schymura, Peng Guo, Yanir Maymon, Boaz Rafaely, Dorothea Kolossa

Department of Electrical & Computer Engineering

Ben-Gurion University of the Negev

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

Abstract

This paper introduces a framework for robust speaker localization in reverberant environments based on a causal analysis of the temporal relationship between direct sound and corresponding reflections. It extends previously proposed localization approaches for spherical microphone arrays based on a direct-path dominance test. So far, these methods are applied in the time-frequency domain without considering the temporal context of direction-of-arrival measurements. In this work, a causal analysis of the temporal structure of subsequent directions-of-arrival estimates based on the Granger causality test is proposed. The cause-effect relationship between estimated directions is modeled via a causal graph, which is used to distinguish the direction of the direct sound from corresponding reflections. An experimental evaluation in simulated acoustic environments shows that the proposed approach yields an improvement in localization performance especially in highly reverberant conditions.

Original language	American English
Title of host publication	Latent Variable Analysis and Signal Separation - 14th International Conference, LVA/ICA 2018, Proceedings
Editors	Sharon Gannot, Yannick Deville, Russell Mason, Mark D. Plumbley, Dominic Ward
Publisher	Springer Verlag
Pages	228-237
Number of pages	10
ISBN (Print)	9783319937632
DOIs	https://doi.org/10.1007/978-3-319-93764-9_22
State	Published - 1 Jan 2018
Event	14th International Conference on Latent Variable Analysis and Signal Separation, LVA/ICA 2018 - Guildford, United Kingdom Duration: 2 Jul 2018 → 5 Jul 2018

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	10891 LNCS

Conference

Conference	14th International Conference on Latent Variable Analysis and Signal Separation, LVA/ICA 2018
Country/Territory	United Kingdom
City	Guildford
Period	2/07/18 → 5/07/18

Keywords

Multivariate granger causality test
Speaker localization
Spherical microphone arrays
Vector autoregressive models

All Science Journal Classification (ASJC) codes

Theoretical Computer Science
General Computer Science

Access to Document

https://doi.org/10.1007/978-3-319-93764-9_22

Cite this

Schymura, C., Guo, P., Maymon, Y., Rafaely, B., & Kolossa, D. (2018). Exploiting structures of temporal causality for robust speaker localization in reverberant environments. In S. Gannot, Y. Deville, R. Mason, M. D. Plumbley, & D. Ward (Eds.), Latent Variable Analysis and Signal Separation - 14th International Conference, LVA/ICA 2018, Proceedings (pp. 228-237). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10891 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-319-93764-9_22

Schymura, Christopher ; Guo, Peng ; Maymon, Yanir et al. / Exploiting structures of temporal causality for robust speaker localization in reverberant environments. Latent Variable Analysis and Signal Separation - 14th International Conference, LVA/ICA 2018, Proceedings. editor / Sharon Gannot ; Yannick Deville ; Russell Mason ; Mark D. Plumbley ; Dominic Ward. Springer Verlag, 2018. pp. 228-237 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{b6d0dc1e6e1e4a7a911e6e81ad1cf825,

title = "Exploiting structures of temporal causality for robust speaker localization in reverberant environments",

abstract = "This paper introduces a framework for robust speaker localization in reverberant environments based on a causal analysis of the temporal relationship between direct sound and corresponding reflections. It extends previously proposed localization approaches for spherical microphone arrays based on a direct-path dominance test. So far, these methods are applied in the time-frequency domain without considering the temporal context of direction-of-arrival measurements. In this work, a causal analysis of the temporal structure of subsequent directions-of-arrival estimates based on the Granger causality test is proposed. The cause-effect relationship between estimated directions is modeled via a causal graph, which is used to distinguish the direction of the direct sound from corresponding reflections. An experimental evaluation in simulated acoustic environments shows that the proposed approach yields an improvement in localization performance especially in highly reverberant conditions.",

keywords = "Multivariate granger causality test, Speaker localization, Spherical microphone arrays, Vector autoregressive models",

author = "Christopher Schymura and Peng Guo and Yanir Maymon and Boaz Rafaely and Dorothea Kolossa",

note = "Publisher Copyright: {\textcopyright} Springer International Publishing AG, part of Springer Nature 2018.; 14th International Conference on Latent Variable Analysis and Signal Separation, LVA/ICA 2018 ; Conference date: 02-07-2018 Through 05-07-2018",

year = "2018",

month = jan,

day = "1",

doi = "https://doi.org/10.1007/978-3-319-93764-9_22",

language = "American English",

isbn = "9783319937632",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Verlag",

pages = "228--237",

editor = "Sharon Gannot and Yannick Deville and Russell Mason and Plumbley, {Mark D.} and Dominic Ward",

booktitle = "Latent Variable Analysis and Signal Separation - 14th International Conference, LVA/ICA 2018, Proceedings",

address = "Germany",

}

Schymura, C, Guo, P, Maymon, Y, Rafaely, B & Kolossa, D 2018, Exploiting structures of temporal causality for robust speaker localization in reverberant environments. in S Gannot, Y Deville, R Mason, MD Plumbley & D Ward (eds), Latent Variable Analysis and Signal Separation - 14th International Conference, LVA/ICA 2018, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 10891 LNCS, Springer Verlag, pp. 228-237, 14th International Conference on Latent Variable Analysis and Signal Separation, LVA/ICA 2018, Guildford, United Kingdom, 2/07/18. https://doi.org/10.1007/978-3-319-93764-9_22

Exploiting structures of temporal causality for robust speaker localization in reverberant environments. / Schymura, Christopher; Guo, Peng; Maymon, Yanir et al.
Latent Variable Analysis and Signal Separation - 14th International Conference, LVA/ICA 2018, Proceedings. ed. / Sharon Gannot; Yannick Deville; Russell Mason; Mark D. Plumbley; Dominic Ward. Springer Verlag, 2018. p. 228-237 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10891 LNCS).

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

TY - GEN

T1 - Exploiting structures of temporal causality for robust speaker localization in reverberant environments

AU - Schymura, Christopher

AU - Guo, Peng

AU - Maymon, Yanir

AU - Rafaely, Boaz

AU - Kolossa, Dorothea

N1 - Publisher Copyright: © Springer International Publishing AG, part of Springer Nature 2018.

PY - 2018/1/1

Y1 - 2018/1/1

N2 - This paper introduces a framework for robust speaker localization in reverberant environments based on a causal analysis of the temporal relationship between direct sound and corresponding reflections. It extends previously proposed localization approaches for spherical microphone arrays based on a direct-path dominance test. So far, these methods are applied in the time-frequency domain without considering the temporal context of direction-of-arrival measurements. In this work, a causal analysis of the temporal structure of subsequent directions-of-arrival estimates based on the Granger causality test is proposed. The cause-effect relationship between estimated directions is modeled via a causal graph, which is used to distinguish the direction of the direct sound from corresponding reflections. An experimental evaluation in simulated acoustic environments shows that the proposed approach yields an improvement in localization performance especially in highly reverberant conditions.

AB - This paper introduces a framework for robust speaker localization in reverberant environments based on a causal analysis of the temporal relationship between direct sound and corresponding reflections. It extends previously proposed localization approaches for spherical microphone arrays based on a direct-path dominance test. So far, these methods are applied in the time-frequency domain without considering the temporal context of direction-of-arrival measurements. In this work, a causal analysis of the temporal structure of subsequent directions-of-arrival estimates based on the Granger causality test is proposed. The cause-effect relationship between estimated directions is modeled via a causal graph, which is used to distinguish the direction of the direct sound from corresponding reflections. An experimental evaluation in simulated acoustic environments shows that the proposed approach yields an improvement in localization performance especially in highly reverberant conditions.

KW - Multivariate granger causality test

KW - Speaker localization

KW - Spherical microphone arrays

KW - Vector autoregressive models

UR - http://www.scopus.com/inward/record.url?scp=85048558016&partnerID=8YFLogxK

U2 - https://doi.org/10.1007/978-3-319-93764-9_22

DO - https://doi.org/10.1007/978-3-319-93764-9_22

M3 - Conference contribution

SN - 9783319937632

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 228

EP - 237

BT - Latent Variable Analysis and Signal Separation - 14th International Conference, LVA/ICA 2018, Proceedings

A2 - Gannot, Sharon

A2 - Deville, Yannick

A2 - Mason, Russell

A2 - Plumbley, Mark D.

A2 - Ward, Dominic

PB - Springer Verlag

T2 - 14th International Conference on Latent Variable Analysis and Signal Separation, LVA/ICA 2018

Y2 - 2 July 2018 through 5 July 2018

ER -

Schymura C, Guo P, Maymon Y, Rafaely B, Kolossa D. Exploiting structures of temporal causality for robust speaker localization in reverberant environments. In Gannot S, Deville Y, Mason R, Plumbley MD, Ward D, editors, Latent Variable Analysis and Signal Separation - 14th International Conference, LVA/ICA 2018, Proceedings. Springer Verlag. 2018. p. 228-237. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: https://doi.org/10.1007/978-3-319-93764-9_22

Exploiting structures of temporal causality for robust speaker localization in reverberant environments

Abstract

Publication series

Conference

Keywords

All Science Journal Classification (ASJC) codes

Access to Document

Other files and links

Fingerprint

Cite this