• JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
 
  Bookmark and Share
 
 
Master's Dissertation
DOI
https://doi.org/10.11606/D.76.2008.tde-11122008-194055
Document
Author
Full name
Paulo César Fantinato
E-mail
Institute/School/College
Knowledge Area
Date of Defense
Published
São Carlos, 2008
Supervisor
Committee
Guido, Rodrigo Capobianco (President)
Martins, Mateus Jose
Pereira, Aledir Silveira
Title in Portuguese
Segmentação de voz baseada na análise fractal e na transformada wavelet.
Keywords in Portuguese
1. Voz. 2. Fractais. 3. Wavelets.
Abstract in Portuguese
Ultimamente, a análise fractal (AF) vem sendo utilizada com sucesso na área de processamento digital de voz, especialmente para fins de segmentação de palavras e fonemas, que é uma das etapas fundamentais dos sistemas de reconhecimento automático de fala (Automatic Speech Recognition - ASR ) e identificação automática de locutor (Automatic Speaker Identification - ASI). O uso prático da AF para ASR e ASI depende de dois fatores básicos: baixo custo computacional, para permitir o uso em tempo-real, e precisão nos resultados, para produzir a segmentação correta e entregar dados coerentes à etapa de classificação. Visando atender a esses objetivos, o presente trabalho propõe uma técnica de segmentação de sinais de voz baseada na dimensão do fractal, obtida com o uso da transformada wavelet discreta (DWT). Diversas famílias de wavelets são testadas e comparadas, sendo que os testes foram realizados com algumas sentenças extraídas da base de dados TIMIT do Linguistic Data Consortium (LDC).
Title in English
Speech segmentation based on fractal analysis and wavelet transform.
Keywords in English
1. Speech 2. Fractals 3. Wavelets
Abstract in English
Nowadays, fractal analysis has been successfully applied to digital speech processing, particularly for words and phonemes segmentation, which represents one of the fundamental steps in automatic speech recognition and speaker identification systems. The practical use of fractal analysis for these purposes should match two principles: low computational cost, to allow use in real-time, and accuracy in the results, to produce a correct segmentation, delivering consistent data to the classifier. Aiming at meeting these two requirements, this work proposes a technique for speech segmentation based on the fractal dimension, obtained by using the discrete wavelet transform (DWT). Many families of wavelets were tested and compared, being the experiments performed with speech data collected from TIMIT corpus provided by the Linguistic Data Consortium.
 
WARNING - Viewing this document is conditioned on your acceptance of the following terms of use:
This document is only for private use for research and teaching activities. Reproduction for commercial use is forbidden. This rights cover the whole data about this document as well as its contents. Any uses or copies of this document in whole or in part must include the author's name.
Publishing Date
2008-12-18
 
WARNING: Learn what derived works are clicking here.
All rights of the thesis/dissertation are from the authors
CeTI-SC/STI
Digital Library of Theses and Dissertations of USP. Copyright © 2001-2024. All rights reserved.