Result Details

Simplification and optimization of I-Vector Extraction

GLEMBEK, O.; BURGET, L.; KENNY, P.; KARAFIÁT, M.; MATĚJKA, P. Simplification and optimization of I-Vector Extraction. Proceedings of the 2011 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2011. Praha: IEEE Signal Processing Society, 2011. p. 4516-4519. ISBN: 978-1-4577-0537-3.
Type
conference paper
Language
English
Authors
Glembek Ondřej, Ing., Ph.D., DCGM (FIT)
Burget Lukáš, doc. Ing., Ph.D., DCGM (FIT)
Kenny Patrick
Karafiát Martin, Ing., Ph.D., DCGM (FIT)
Matějka Pavel, Ing., Ph.D., DCGM (FIT)
Abstract

We managed to reduce the memory requirements and processing time for the i-vector extractor training so that higher dimensions can be now used while retaining the recognition accuracy. As for i-vector extraction, we managed to reduce the complexity of the algorithm with sacrificing little recognition accuracy, which makes this technique usable in small-scale devices.

Keywords

speaker recognition, i-vectors, Joint Factor Analysis, PCA, HLDA

URL
Annotation

This paper introduces some simplifications to the i-vector speaker recognition systems. I-vector extraction as well as training of the i-vector extractor can be an expensive task both in terms of memory and speed. Under certain assumptions, the formulas for i-vector extraction-also used in i-vector extractor training-can be simplified and lead to a faster and memory more efficient code. The first assumption is that the GMM component alignment is constant across utterances and is given by the UBM GMM weights. The second assumption is that the i-vector extractor matrix can be linearly transformed so that its per-Gaussian components are orthogonal. We use PCA and HLDA to estimate this transform.

Published
2011
Pages
4516–4519
Proceedings
Proceedings of the 2011 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2011
Conference
International Conference on Acoustics, Speech and Signal Processing 2011
ISBN
978-1-4577-0537-3
Publisher
IEEE Signal Processing Society
Place
Praha
BibTeX 
@inproceedings{BUT76376,
  author="Ondřej {Glembek} and Lukáš {Burget} and Patrick {Kenny} and Martin {Karafiát} and Pavel {Matějka}",
  title="Simplification and optimization of I-Vector Extraction",
  booktitle="Proceedings of the 2011 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2011",
  year="2011",
  pages="4516--4519",
  publisher="IEEE Signal Processing Society",
  address="Praha",
  isbn="978-1-4577-0537-3",
  url="http://www.fit.vutbr.cz/research/groups/speech/publi/2011/glembek_icassp2011_4516.pdf"
}
Projects
Mobile Biometry, MŠMT, Podpora projektů sedmého rámcového programu Evropského společenství pro výzkum, technologický rozvoj a demonstrace (2007 až 2013) podle zákona č. 171/2007 Sb., 7E08042, start: 2008-01-01, end: 2010-12-31, completed
Recognition and presentation of multimedia data, BUT, Vnitřní projekty VUT, FIT-S-10-2, 2010, start: 2010-04-01, end: 2010-12-31, completed
Security-Oriented Research in Information Technology, MŠMT, Institucionální prostředky SR ČR (např. VZ, VC), MSM0021630528, start: 2007-01-01, end: 2013-12-31, running
Speech Recognition under Real-World Conditions, GACR, Standardní projekty, GA102/08/0707, start: 2008-01-01, end: 2011-12-31, completed
Research groups
Speech Data Mining Research Group BUT Speech@FIT (RG SPEECH)
Departments
Department of Computer Graphics and Multimedia (DCGM)
Back to top