Title: A direct criterion minimization based fMLLR via gradient descend
Authors: Vaněk, Jan
Zajíc, Zbyněk
Citation: VANĚK, Jan; ZAJÍC, Zbyněk. A direct criterion minimization based fMLLR via gradient descend. In: Text, speech and dialogue. Berlin: Springer, 2013, p. 52-59. (Lectures notes in computer science; 8082). ISBN 978-3-642-40584-6.
Issue Date: 2013
Publisher: Springer
Document type: článek
URI: http://www.kky.zcu.cz/cs/publications/JanVanek_2013_ADirectCriterion
ISBN: 978-3-642-40584-6
Keywords: ASR;fMLLR;adaptace
Keywords in different language: ASR;fMLLR;adaptation
Abstract in different language: Adaptation techniques are necessary in automatic speech recognizers to improve a recognition accuracy. Linear Transformation methods (MLLR or fMLLR) are the most favorite in the case of limited available data. The fMLLR is the feature-space transformation. This is the advantage with contrast to MLLR that transforms the entire acoustic model. The classical fMLLR estimation involves maximization of the likelihood criterion based on individual Gaussian components statistic.We proposed an approach which takes into account the overall likelihood of a HMMstate. It estimates the transformation to optimize the ML criterion of HMM directly using gradient descent algorithm.
Rights: © Jan Vaněk - Zbyněk Zajíc
Appears in Collections:Články / Articles (NTIS)

Files in This Item:
File Description SizeFormat 
JanVanek_2013_ADirectCriterion.pdfPlný text101,06 kBAdobe PDFView/Open

Please use this identifier to cite or link to this item: http://hdl.handle.net/11025/17162

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