Developing a Speaker Identification System for the DARPA RATS Project

O Plchot; S Matsoukas; P Matejka; N Dehak; J Ma; S. Cumani; O Glembek; H Hermansky; SH Mallidi; N Mesgarani; R Schwartz; M Soufifar; Zheng-Hua Tan; S Thomas; B Zhang; X Zhou

doi:10.1109/ICASSP.2013.6638972

Developing a Speaker Identification System for the DARPA RATS Project

O Plchot, S Matsoukas, P Matejka, N Dehak, J Ma, S. Cumani, O Glembek, H Hermansky, SH Mallidi, N Mesgarani, R Schwartz, M Soufifar, Zheng-Hua Tan, S Thomas, B Zhang, X Zhou

Institut for Elektroniske Systemer

Publikation: Bidrag til bog/antologi/rapport/konference proceeding › Konferenceartikel i proceeding › Forskning › peer review

36 Citationer (Scopus)

Abstract

This paper describes the speaker identification (SID) system developed
by the Patrol team for the first phase of the DARPA RATS (Robust
Automatic Transcription of Speech) program, which seeks to
advance state of the art detection capabilities on audio from highly
degraded communication channels. We present results using multiple
SID systems differing mainly in the algorithm used for voice
activity detection (VAD) and feature extraction. We show that (a)
unsupervised VAD performs as well supervised methods in terms
of downstream SID performance, (b) noise-robust feature extraction
methods such as CFCCs out-perform MFCC front-ends on noisy audio,
and (c) fusion of multiple systems provides 24% relative improvement
in EER compared to the single best system when using a
novel SVM-based fusion algorithm that uses side information such
as gender, language, and channel id.

Originalsprog	Engelsk
Titel	Acoustics, Speech and Signal Processing (ICASSP), 2013 IEEE International Conference on
Antal sider	5
Forlag	IEEE
Publikationsdato	2013
Sider	6768 - 6772
ISBN (Trykt)	978-1-4799-0356-6
DOI	https://doi.org/10.1109/ICASSP.2013.6638972
Status	Udgivet - 2013
Begivenhed	2013 IEEE International Conference on Acoustics, Speech, and Signal Processing - Vancouver, Canada Varighed: 26 maj 2013 → 31 maj 2013 Konferencens nummer: 38

Konference

Konference	2013 IEEE International Conference on Acoustics, Speech, and Signal Processing
Nummer	38
Land/Område	Canada
By	Vancouver
Periode	26/05/2013 → 31/05/2013

Navn	I E E E International Conference on Acoustics, Speech and Signal Processing. Proceedings
ISSN	1520-6149

Adgang til dokumentet

10.1109/ICASSP.2013.6638972

AUB Link

Søg efter materialet i Aalborg Universitetsbiblioteks søgemaskine

Citationsformater

Plchot, O., Matsoukas, S., Matejka, P., Dehak, N., Ma, J., Cumani, S., Glembek, O., Hermansky, H., Mallidi, SH., Mesgarani, N., Schwartz, R., Soufifar, M., Tan, Z.-H., Thomas, S., Zhang, B., & Zhou, X. (2013). Developing a Speaker Identification System for the DARPA RATS Project. I Acoustics, Speech and Signal Processing (ICASSP), 2013 IEEE International Conference on (s. 6768 - 6772). IEEE. https://doi.org/10.1109/ICASSP.2013.6638972

@inproceedings{474b16501fbc44f48ffe72293bd6fc09,

title = "Developing a Speaker Identification System for the DARPA RATS Project",

abstract = "This paper describes the speaker identification (SID) system developedby the Patrol team for the first phase of the DARPA RATS (RobustAutomatic Transcription of Speech) program, which seeks toadvance state of the art detection capabilities on audio from highlydegraded communication channels. We present results using multipleSID systems differing mainly in the algorithm used for voiceactivity detection (VAD) and feature extraction. We show that (a)unsupervised VAD performs as well supervised methods in termsof downstream SID performance, (b) noise-robust feature extractionmethods such as CFCCs out-perform MFCC front-ends on noisy audio,and (c) fusion of multiple systems provides 24% relative improvementin EER compared to the single best system when using anovel SVM-based fusion algorithm that uses side information suchas gender, language, and channel id.",

author = "O Plchot and S Matsoukas and P Matejka and N Dehak and J Ma and S. Cumani and O Glembek and H Hermansky and SH Mallidi and N Mesgarani and R Schwartz and M Soufifar and Zheng-Hua Tan and S Thomas and B Zhang and X Zhou",

year = "2013",

doi = "10.1109/ICASSP.2013.6638972",

language = "English",

isbn = "978-1-4799-0356-6",

series = "I E E E International Conference on Acoustics, Speech and Signal Processing. Proceedings",

publisher = "IEEE",

pages = "6768 -- 6772",

booktitle = "Acoustics, Speech and Signal Processing (ICASSP), 2013 IEEE International Conference on",

address = "United States",

note = "2013 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP ; Conference date: 26-05-2013 Through 31-05-2013",

}

Plchot, O, Matsoukas, S, Matejka, P, Dehak, N, Ma, J, Cumani, S, Glembek, O, Hermansky, H, Mallidi, SH, Mesgarani, N, Schwartz, R, Soufifar, M, Tan, Z-H, Thomas, S, Zhang, B & Zhou, X 2013, Developing a Speaker Identification System for the DARPA RATS Project. i Acoustics, Speech and Signal Processing (ICASSP), 2013 IEEE International Conference on . IEEE, I E E E International Conference on Acoustics, Speech and Signal Processing. Proceedings, s. 6768 - 6772, 2013 IEEE International Conference on Acoustics, Speech, and Signal Processing, Vancouver, Canada, 26/05/2013. https://doi.org/10.1109/ICASSP.2013.6638972

Developing a Speaker Identification System for the DARPA RATS Project. / Plchot, O; Matsoukas, S; Matejka, P et al.
Acoustics, Speech and Signal Processing (ICASSP), 2013 IEEE International Conference on . IEEE, 2013. s. 6768 - 6772 (I E E E International Conference on Acoustics, Speech and Signal Processing. Proceedings).

Publikation: Bidrag til bog/antologi/rapport/konference proceeding › Konferenceartikel i proceeding › Forskning › peer review

TY - GEN

T1 - Developing a Speaker Identification System for the DARPA RATS Project

AU - Plchot, O

AU - Matsoukas, S

AU - Matejka, P

AU - Dehak, N

AU - Ma, J

AU - Cumani, S.

AU - Glembek, O

AU - Hermansky, H

AU - Mallidi, SH

AU - Mesgarani, N

AU - Schwartz, R

AU - Soufifar, M

AU - Tan, Zheng-Hua

AU - Thomas, S

AU - Zhang, B

AU - Zhou, X

N1 - Conference code: 38

PY - 2013

Y1 - 2013

N2 - This paper describes the speaker identification (SID) system developedby the Patrol team for the first phase of the DARPA RATS (RobustAutomatic Transcription of Speech) program, which seeks toadvance state of the art detection capabilities on audio from highlydegraded communication channels. We present results using multipleSID systems differing mainly in the algorithm used for voiceactivity detection (VAD) and feature extraction. We show that (a)unsupervised VAD performs as well supervised methods in termsof downstream SID performance, (b) noise-robust feature extractionmethods such as CFCCs out-perform MFCC front-ends on noisy audio,and (c) fusion of multiple systems provides 24% relative improvementin EER compared to the single best system when using anovel SVM-based fusion algorithm that uses side information suchas gender, language, and channel id.

AB - This paper describes the speaker identification (SID) system developedby the Patrol team for the first phase of the DARPA RATS (RobustAutomatic Transcription of Speech) program, which seeks toadvance state of the art detection capabilities on audio from highlydegraded communication channels. We present results using multipleSID systems differing mainly in the algorithm used for voiceactivity detection (VAD) and feature extraction. We show that (a)unsupervised VAD performs as well supervised methods in termsof downstream SID performance, (b) noise-robust feature extractionmethods such as CFCCs out-perform MFCC front-ends on noisy audio,and (c) fusion of multiple systems provides 24% relative improvementin EER compared to the single best system when using anovel SVM-based fusion algorithm that uses side information suchas gender, language, and channel id.

U2 - 10.1109/ICASSP.2013.6638972

DO - 10.1109/ICASSP.2013.6638972

M3 - Article in proceeding

SN - 978-1-4799-0356-6

T3 - I E E E International Conference on Acoustics, Speech and Signal Processing. Proceedings

SP - 6768

EP - 6772

BT - Acoustics, Speech and Signal Processing (ICASSP), 2013 IEEE International Conference on

PB - IEEE

T2 - 2013 IEEE International Conference on Acoustics, Speech, and Signal Processing

Y2 - 26 May 2013 through 31 May 2013

ER -

Plchot O, Matsoukas S, Matejka P, Dehak N, Ma J, Cumani S et al. Developing a Speaker Identification System for the DARPA RATS Project. I Acoustics, Speech and Signal Processing (ICASSP), 2013 IEEE International Conference on . IEEE. 2013. s. 6768 - 6772. (I E E E International Conference on Acoustics, Speech and Signal Processing. Proceedings). doi: 10.1109/ICASSP.2013.6638972

Developing a Speaker Identification System for the DARPA RATS Project

Abstract

Konference

Adgang til dokumentet

AUB Link

Fingeraftryk

Citationsformater