TEDxSK and JumpSK: A New Slovak Speech Recognition Dedicated Corpus-Reference-Cited by-同舟云学术

TEDxSK and JumpSK: A New Slovak Speech Recognition Dedicated Corpus

Published:2017-12-01 Issue:2 Volume:68 Page:346-354
ISSN:1338-4287
Container-title:Journal of Linguistics/Jazykovedný casopis
language:en
Short-container-title:

Author:

Staš Ján¹,Hládek Daniel¹,Viszlay Peter¹,Koctúr Tomáš¹

Affiliation:

1. Faculty of Electrical Engineering and Informatics , Technical University of Košice , Slovakia

Abstract

Abstract This paper describes a new Slovak speech recognition dedicated corpus built from TEDx talks and Jump Slovakia lectures. The proposed speech database consists of 220 talks and lectures in total duration of about 58 hours. Annotated speech database was generated automatically in an unsupervised manner by using acoustic speech segmentation based on principal component analysis and automatic speech transcription using two complementary speech recognition systems. The evaluation data consisting of 50 manually annotated talks and lectures in total duration of about 12 hours, has been created for evaluation of the quality of Slovak speech recognition. By unsupervised automatic annotation of TEDx talks and Jump Slovakia lectures we have obtained 21.26% of new speech segments with approximately 9.44% word error rate, suitable for retraining or adaptation of acoustic models trained beforehand.

Publisher

Walter de Gruyter GmbH

Subject

Linguistics and Language,Language and Linguistics

Reference28 articles.

1. [1] Koctúr, T., Juhár, J., Viszlay, P., Staš, J., and Lojka, M. (2016). Unsupervised speech transcription and alignment based on two complementary ASR systems. In Proceedings of RADIOELEKTRONIKA 2016, pages 358–362, Košice, Slovakia.

2. [2] Rosseau, A., Deléglise, P., and Estève, Y. (2012). TED-LIUM: An automatic speech recognition dedicated corpus. In Proceedings of LREC 2012, pages 125–129, Istanbul, Turkey.

3. [3] Deléglise, P., Estève, Y., Meignier, S., and Merlin, T. (2009). Improvements to the LIUM French ASR system based on CMU Sphinx: What helps to significantly reduce the word error rate? In Proceedings of INTERSPEECH 2009, pages 2123–2126, Brighton, UK.

4. [4] Žgank, A., Maučec, M. S., Verdonik, D. (2016). The SI TEDx-UM speech database: A new Slovenian spoken language resource. In Proceedings of LREC 2016, pages 4670–4673, Portorož, Slovenia.

5. [5] Rosseau, A., Deléglise, P., and Estève, Y. (2014). Enhancing the TED-LIUM corpus with selected data for language modeling and more TED talks. In Proceedings of LREC 2014, pages 3935–3939, Reykjavik, Iceland.