Publication Details

Discriminatively Re-trained i-Vector Extractor For Speaker Recognition

NOVOTNÝ Ondřej, PLCHOT Oldřich, GLEMBEK Ondřej, BURGET Lukáš and MATĚJKA Pavel. Discriminatively Re-trained i-Vector Extractor For Speaker Recognition. In: Proceedings of ICASSP. Brighton: IEEE Signal Processing Society, 2019, pp. 6031-6035. ISBN 978-1-5386-4658-8. Available from: https://ieeexplore.ieee.org/document/8682590
Czech title
Diskriminativně přetrénovaný extraktor i-vektorů pro rozpoznávání mluvčího
Type
conference paper
Language
english
Authors
URL
Keywords
i-vectors, i-vector extractor, speaker recogni-tion, speaker verification, discriminative training
Abstract
In this work we revisit discriminative training of the i-vector extractor component in the standard speaker verification (SV) system. The motivation of our research lies in the robustness and stability of this large generative model, which we want to preserve, and focus its power towards any intended SV task. We show that after generative initialization of the i-vector extractor, we can further refine it with discriminative training and obtain i-vectors that lead to better performance on various benchmarks representing different acoustic domains.
Published
2019
Pages
6031-6035
Proceedings
Proceedings of ICASSP
Conference
International Conference on Acoustics, Speech, and Signal Processing, Brighton, GB
ISBN
978-1-5386-4658-8
Publisher
IEEE Signal Processing Society
Place
Brighton, GB
DOI
BibTeX
@INPROCEEDINGS{FITPUB12093,
   author = "Ond\v{r}ej Novotn\'{y} and Old\v{r}ich Plchot and Ond\v{r}ej Glembek and Luk\'{a}\v{s} Burget and Pavel Mat\v{e}jka",
   title = "Discriminatively Re-trained i-Vector Extractor For Speaker Recognition",
   pages = "6031--6035",
   booktitle = "Proceedings of ICASSP",
   year = 2019,
   location = "Brighton, GB",
   publisher = "IEEE Signal Processing Society",
   ISBN = "978-1-5386-4658-8",
   doi = "10.1109/ICASSP.2019.8682590",
   language = "english",
   url = "https://www.fit.vut.cz/research/publication/12093"
}
Back to top