BDBComp
Parceria:
SBC
Time-frequency Analysis and Artificial Intelligence in Pathological Voice Signals Processing

Everthon S. FonsecaAndré C. SilvestreWalison J. BarberaJosé C. PereiraAdilson Gonzaga

An algorithm able to classify pathological and normal voice signals based on DaubechiesDiscrete Wavelet Transform (DWT-db) and Support Vector Machines (SVM) classifier is presented.DWT-db is used for time-frequency analysis giving quantitative evaluation of signal characteristics toidentify pathologies in voice signals, mainly the nodule pathology in vocal folds, of subjects withdifferent ages for both male and female. The signals mean square values of a particular scale fromwavelet analysis are entries to a Least Square Support Vector Machine (LS-SVM) classifier. A nonlinearLS-SVM leads to an adequate larynx pathology classifier (over 90% of classification accuracy)as far as nodule is concerned.

http://www.lbd.dcc.ufmg.br/colecoes/wvc/2005/0025.pdf

Caso o link acima esteja inválido, faça uma busca pelo texto completo na Web: Buscar na Web

Biblioteca Digital Brasileira de Computação - Contato: bdbcomp@lbd.dcc.ufmg.br
     Mantida por:
LBD