BUT OpenSAT 2017 Speech Recognition System

Martin Karafiát, Murali Karthick Baskar, Igor Szöke, Vladimír Malenovský, Karel Veselý, František Grézl, Lukáš Burget, Jan Černocký


The paper describes BUT Automatic Speech Recognition (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.


 DOI: 10.21437/Interspeech.2018-2457

Cite as: Karafiát, M., Baskar, M.K., Szöke, I., Malenovský, V., Veselý, K., Grézl, F., Burget, L., Černocký, J. (2018) BUT OpenSAT 2017 Speech Recognition System. Proc. Interspeech 2018, 2638-2642, DOI: 10.21437/Interspeech.2018-2457.


@inproceedings{Karafiát2018,
  author={Martin Karafiát and Murali Karthick Baskar and Igor Szöke and Vladimír Malenovský and Karel Veselý and František Grézl and Lukáš Burget and Jan Černocký},
  title={BUT OpenSAT 2017 Speech Recognition System},
  year=2018,
  booktitle={Proc. Interspeech 2018},
  pages={2638--2642},
  doi={10.21437/Interspeech.2018-2457},
  url={http://dx.doi.org/10.21437/Interspeech.2018-2457}
}