Přístupnostní navigace
E-application
Search Search Close
Publication detail
KARAFIÁT, M. BASKAR, M. SZŐKE, I. MALENOVSKÝ, V. VESELÝ, K. GRÉZL, F. BURGET, L. ČERNOCKÝ, J.
Original Title
BUT OpenSAT 2017 speech recognition system
Type
conference paper
Language
English
Original Abstract
(ASR) systems for two domains in OpenSAT evaluations: Low Resourced Languages and Public Safety Communications. The first was challenging due to lack of training data, therefore multilingual approaches for BLSTM training were employed and recently published Residual Memory Networks requiring less training data were used. Combination of both approaches led to superior performance. The second domain was challenging due to recording in extreme conditions: specific channel, speaker under stress, high levels of noise. A data augmentation process was very important to get reasonably good performance.
Keywords
speech recognition, multilingual training, BLSTM, data augmentation, robustness
Authors
KARAFIÁT, M.; BASKAR, M.; SZŐKE, I.; MALENOVSKÝ, V.; VESELÝ, K.; GRÉZL, F.; BURGET, L.; ČERNOCKÝ, J.
Released
2. 9. 2018
Publisher
International Speech Communication Association
Location
Hyderabad
ISBN
1990-9772
Periodical
Proceedings of Interspeech
Year of study
2018
Number
9
State
French Republic
Pages from
2638
Pages to
2642
Pages count
5
URL
https://www.isca-speech.org/archive/Interspeech_2018/abstracts/2457.html
BibTex
@inproceedings{BUT155099, author="Martin {Karafiát} and Murali Karthick {Baskar} and Igor {Szőke} and Vladimír {Malenovský} and Karel {Veselý} and František {Grézl} and Lukáš {Burget} and Jan {Černocký}", title="BUT OpenSAT 2017 speech recognition system", booktitle="Proceedings of Interspeech 2018", year="2018", journal="Proceedings of Interspeech", volume="2018", number="9", pages="2638--2642", publisher="International Speech Communication Association", address="Hyderabad", doi="10.21437/Interspeech.2018-2457", issn="1990-9772", url="https://www.isca-speech.org/archive/Interspeech_2018/abstracts/2457.html" }