Joint Single-Channel Speech Separation and Speaker Identification

Pejman Mowlaee, Rahim Saeidi, Zheng-Hua Tan, Mads Græsbøll Christensen, Pasi Fränti, Søren Holdt Jensen

Publikation: Bidrag til tidsskriftKonferenceartikel i tidsskriftForskningpeer review

16 Citationer (Scopus)
618 Downloads (Pure)

Resumé

In this paper, we propose a closed loop system to improve the performance
of single-channel speech separation in a speaker independent
scenario. The system is composed of two interconnected blocks: a
separation block and a speaker identiſcation block. The improvement
is accomplished by incorporating the speaker identities found
by the speaker identiſcation block as additional information for the
separation block, which converts the speaker-independent separation
problem to a speaker-dependent one where the speaker codebooks
are known. Simulation results show that the closed loop system enhances
the quality of the separated output signals. To assess the improvements,
the results are reported in terms of PESQ for both target
and masked signals.
OriginalsprogEngelsk
TidsskriftI E E E International Conference on Acoustics, Speech and Signal Processing. Proceedings
Vol/bind2010
Sider (fra-til)4430 - 4433
ISSN1520-6149
DOI
StatusUdgivet - 2010
Begivenhed2010 IEEE International Conference on Acoustics, Speech, and Signal Processing - Dallas, USA
Varighed: 14 mar. 201017 mar. 2010

Konference

Konference2010 IEEE International Conference on Acoustics, Speech, and Signal Processing
LandUSA
ByDallas
Periode14/03/201017/03/2010

Fingerprint

Closed loop systems
Identification (control systems)

Citer dette

@inproceedings{59729371a9224edc930b925e1074df9c,
title = "Joint Single-Channel Speech Separation and Speaker Identification",
abstract = "In this paper, we propose a closed loop system to improve the performanceof single-channel speech separation in a speaker independentscenario. The system is composed of two interconnected blocks: aseparation block and a speaker identiſcation block. The improvementis accomplished by incorporating the speaker identities foundby the speaker identiſcation block as additional information for theseparation block, which converts the speaker-independent separationproblem to a speaker-dependent one where the speaker codebooksare known. Simulation results show that the closed loop system enhancesthe quality of the separated output signals. To assess the improvements,the results are reported in terms of PESQ for both targetand masked signals.",
keywords = "Single-channel speech separation, speaker identification, sinusoidal mixture estimator, vector quantization, Gaussian mixture model",
author = "Pejman Mowlaee and Rahim Saeidi and Zheng-Hua Tan and Christensen, {Mads Gr{\ae}sb{\o}ll} and Pasi Fr{\"a}nti and Jensen, {S{\o}ren Holdt}",
year = "2010",
doi = "10.1109/ICASSP.2010.5495619",
language = "English",
volume = "2010",
pages = "4430 -- 4433",
journal = "I E E E International Conference on Acoustics, Speech and Signal Processing. Proceedings",
issn = "1520-6149",
publisher = "IEEE Signal Processing Society",

}

Joint Single-Channel Speech Separation and Speaker Identification. / Mowlaee, Pejman; Saeidi, Rahim ; Tan, Zheng-Hua; Christensen, Mads Græsbøll; Fränti, Pasi ; Jensen, Søren Holdt.

I: I E E E International Conference on Acoustics, Speech and Signal Processing. Proceedings, Bind 2010, 2010, s. 4430 - 4433 .

Publikation: Bidrag til tidsskriftKonferenceartikel i tidsskriftForskningpeer review

TY - GEN

T1 - Joint Single-Channel Speech Separation and Speaker Identification

AU - Mowlaee, Pejman

AU - Saeidi, Rahim

AU - Tan, Zheng-Hua

AU - Christensen, Mads Græsbøll

AU - Fränti, Pasi

AU - Jensen, Søren Holdt

PY - 2010

Y1 - 2010

N2 - In this paper, we propose a closed loop system to improve the performanceof single-channel speech separation in a speaker independentscenario. The system is composed of two interconnected blocks: aseparation block and a speaker identiſcation block. The improvementis accomplished by incorporating the speaker identities foundby the speaker identiſcation block as additional information for theseparation block, which converts the speaker-independent separationproblem to a speaker-dependent one where the speaker codebooksare known. Simulation results show that the closed loop system enhancesthe quality of the separated output signals. To assess the improvements,the results are reported in terms of PESQ for both targetand masked signals.

AB - In this paper, we propose a closed loop system to improve the performanceof single-channel speech separation in a speaker independentscenario. The system is composed of two interconnected blocks: aseparation block and a speaker identiſcation block. The improvementis accomplished by incorporating the speaker identities foundby the speaker identiſcation block as additional information for theseparation block, which converts the speaker-independent separationproblem to a speaker-dependent one where the speaker codebooksare known. Simulation results show that the closed loop system enhancesthe quality of the separated output signals. To assess the improvements,the results are reported in terms of PESQ for both targetand masked signals.

KW - Single-channel speech separation

KW - speaker identification

KW - sinusoidal mixture estimator

KW - vector quantization

KW - Gaussian mixture model

U2 - 10.1109/ICASSP.2010.5495619

DO - 10.1109/ICASSP.2010.5495619

M3 - Conference article in Journal

VL - 2010

SP - 4430

EP - 4433

JO - I E E E International Conference on Acoustics, Speech and Signal Processing. Proceedings

JF - I E E E International Conference on Acoustics, Speech and Signal Processing. Proceedings

SN - 1520-6149

ER -