Skip navigation
Please use this identifier to cite or link to this item: http://uvadoc.uva.es/handle/10324/27529
Title: Evaluation different non-native pronunciation scoring metrics with the Japanese speakers of the sample corpus
Authors: Álvarez-Álvarez, Vandria
Escudero-Mancebo, David
González-Ferreras, César
Cardeñoso-Payo, Valentín
Conference: Iberspeech: "IX Jornadas en Tecnologías del Habla" y "V Iberian Sltech Workshop"
Issue Date: 2016
Abstract: This work presents an analysis over the set of results derived from the goodness of pronunciation (GOP) algorithm for the evaluation of pronunciation at phoneme level over the SAMPLE corpus of non native speech. This corpus includes several recordings of uttered sentences by distinct speakers that have been rated in terms of quality by a group of linguists. The utterances have been automatically rated with the GOP algorithm. The phoneme dependence is discussed to suggest the normalization of intermediate results that could enhance the metrics performance. As result, new scoring proposals are presented which are based on computing the log-likelihood values obtained from the GOP algorithm and the application of a set of rules. These new scores show to correlate with the human rates better than the original GOP metric.
Sponsor: MEC-FEDER Grant TIN2014-59852-R y la Junta de Castilla y León Regional Grant VA145U14
Rights Owner: Springer International Publishing AG
Language: eng
URI: http://uvadoc.uva.es/handle/10324/27529
Rights: info:eu-repo/semantics/openAccess
Appears in Collections:ECA-SIMM - Comunicaciones a congresos, conferencias, etc.

Files in This Item:
File Description SizeFormat 
alvarez16.pdf325,41 kBAdobe PDFThumbnail
View/Open

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

Suggestions
University of Valladolid
Powered by MIT's. DSpace software, Version 5.5
UVa-STIC