Title:	LSTM-Based Speech Segmentation Trained on Different Foreign Languages
Authors:	Hanzlíček, Zdeněk Vít, Jakub
Citation:	HANZLÍČEK, Z. VÍT, J. LSTM-Based Speech Segmentation Trained on Different Foreign Languages. In: Text, Speech, and Dialogue 23rd International Conference, TSD 2020, Brno, Czech Republic, September 8-11, 2020, Proceedings. Cham: Springer Nature Switzerland AG, 2020. s. 456-464. ISBN 978-3-030-58322-4, ISSN 0302-9743.
Issue Date:	2020
Publisher:	Springer Nature Switzerland AG
Document type:	konferenční příspěvek conferenceObject
URI:	2-s2.0-85091145791 http://hdl.handle.net/11025/43117
ISBN:	978-3-030-58322-4
ISSN:	0302-9743
Keywords in different language:	Speech segmentation;Neural networks;LSTM
Abstract in different language:	This paper describes experiments on speech segmentation by using bidirectional LSTM neural networks. The networks were trained on various languages (English, German, Russian and Czech), segmentation experiments were performed on 4 Czech professional voices. To be able to use various combinations of foreign languages, we defined a reduced phonetic alphabet based on IPA notation. It consists of 26 phones, all included in all languages. To increase the segmentation accuracy, we applied an iterative procedure based on detection of improperly segmented data and retraining of the network. Experiments confirmed the convergence of the procedure. A comparison with a reference HMM-based segmentation with additional manual corrections was performed.
Rights:	Plný text není přístupný. © Springer
Appears in Collections:	Konferenční příspěvky / Conference papers (NTIS) Konferenční příspěvky / Conference Papers (KKY) OBD

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.