Другие журналы

. A B C D E F G H I J K L M N O P R S T U V W Y Z К М С Т

Akat'ev

Adaptive cluster model of minimal speech units in analysis and speech recognition problems
Engineering Education # 02, February 2013
DOI: 10.7463/0213.0527867
This article deals with the problem of variability of word pronunciation in analysis and speech recognition tasks. An adaptive acoustic model defined as a multitude of minimal sound units (MSU) united into a cluster-phoneme under the principle of minimum informational mismatch in Kullback-Leibler metric, is proposed. An adaptive algorithm of filling the MSU cluster from a continuous stream of speech was developed on the basis of the whitening filter method. An example of its practical implementation is also provided in the article. As a result of this experiment, from the total list of phonemes of the national language the authors selected the phonemes which, in their implementation, are the most sensitive to conditions of their pronunciation by the speaker. Adjusting an information system to such a phoneme, the authors guarantee maximum sensitivity of perception in relation to the speaker’s emotional and physical state.
 
SEARCH
 
elibrary crossref ulrichsweb neicon rusycon
Photos
 
Events
 
News



Authors
Press-releases
Library
Conferences
About Project
Rambler's Top100
Phone: +7 (915) 336-07-65 (строго: среда; пятница c 11-00 до 17-00)
  RSS
© 2003-2020 «Наука и образование»
Перепечатка материалов журнала без согласования с редакцией запрещена
 Phone: +7 (915) 336-07-65 (строго: среда; пятница c 11-00 до 17-00)