Title: Annotation errors detection in TTS corpora
Other Titles: Detekce anotaÄŤnĂ­ch chyb v korpusech pro TTS
Authors: Matoušek, Jindřich
Tihelka, Daniel
Citation: MATOUĹ EK, JindĹ™ich; TIHELKA, Daniel. Annotation errors detection in TTS corpora. In: Proceedings of ICSPL 2013: 14th Annual Conference of the International Speech Communication Association 2013, Lyon, France, 25-29 August 2013. [Baixas]: ISCA, 2013, p. 1511-1515. IBSN 978-1-62993-443-3.
Issue Date: 2013
Publisher: ISCA
Document type: článek
article
URI: http://www.kky.zcu.cz/cs/publications/MatousekJ_2013_AnnotationErrors
http://hdl.handle.net/11025/17010
ISBN: 978-1-62993-443-3
Keywords: detekce anotaÄŤnĂ­ch chyb;klasifikace;detekce novinek;syntĂ©za Ĺ™eÄŤi
Keywords in different language: annotation error detection;classification;novelty detection;speech synthesis
Abstract in different language: We investigate the problem of automatic detection of annotation errors in single-speaker read-speech corpora used for text-to-speech (TTS) synthesis. Various word-level feature sets were used, and the performance of several detection methods based on support vector machines, extremely randomized trees, k-nearest neighbors, and the performance of novelty and outlier detection are evaluated. We show that both word- and utterance-level annotation error detections perform very well with both high precision and recall scores and with F1 measure being almost 90%, or 97%, respectively.
Rights: Â© JindĹ™ich Matoušek - Daniel Tihelka
Appears in Collections:Články / Articles (NTIS)
Články / Articles (KIV)

Files in This Item:
File Description SizeFormat 
MatousekJ_2013_AnnotationErrors.pdfPlný text214,82 kBAdobe PDFView/Open


Please use this identifier to cite or link to this item: http://hdl.handle.net/11025/17010

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