Respiratory sounds classification using Gaussian mixture models

Author

Bahoura, Mohammed ; Pelletier, Charles

Author_Institution

DMIG, Univ. du Quebec a Rimouski, Que., Canada

Volume

3

fYear

2004

Firstpage

1309

Abstract

The Gaussian mixture models (GMM) method is proposed to classify respiratory sounds in two categories: normal and wheezing. The sound signal is divided in overlapped segments, which are characterized by reduced dimension feature vectors using cepstral or wavelet transforms. The proposed method is compared with other classifiers: vector quantization (VQ) and multilayer perceptron (MLP) neural networks. A post processing is proposed to improve the test results.

Keywords

Gaussian processes; acoustic signal processing; medical signal processing; multilayer perceptrons; pattern classification; signal classification; vector quantisation; wavelet transforms; Gaussian mixture models; MLP; VQ; cepstral transforms; multilayer perceptron neural networks; overlapped segments; reduced dimension feature vectors; respiratory sounds classification; vector quantization; wavelet transforms; wheezing; Cepstral analysis; Covariance matrix; Feature extraction; Fourier transforms; Maximum likelihood estimation; Multilayer perceptrons; Neural networks; Testing; Vector quantization; Wavelet transforms;

fLanguage

English

Publisher

ieee

Conference_Titel

Electrical and Computer Engineering, 2004. Canadian Conference on

ISSN

0840-7789

Print_ISBN

0-7803-8253-6

Type

conf

DOI

10.1109/CCECE.2004.1349639

Filename

1349639