Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Respiratory sounds classification using cepstral analysis and Gaussian mixture models.

M Bahoura1, C Pelletier

  • 1Département de Mathématiques, d'Informatique et de Génie, Université du Québec à Rimouski, Que., Canada.

Conference Proceedings : ... Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Annual Conference
|February 3, 2007
PubMed
Summary

This study introduces a Gaussian Mixture Model (GMM) approach for classifying respiratory sounds, distinguishing normal breathing from wheezing using cepstral analysis and feature extraction methods.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Report summary - Mood and Anxiety Disorders in Canada, 2016.

Health promotion and chronic disease prevention in Canada : research, policy and practice·2016
Same author

Chromatographic, NMR and vibrational spectroscopic investigations of astaxanthin esters: application to "Astaxanthin-rich shrimp oil" obtained from processing of Nordic shrimps.

The Analyst·2015
Same author

Extreme tunability in aluminum doped zinc oxide plasmonic materials for near-infrared applications.

Scientific reports·2014
Same author

[Anaesthetic management of a parturient with a severe factor VII deficiency: was activated factor VII really necessary?].

Annales francaises d'anesthesie et de reanimation·2014
Same author

What about running on Mars? Dynamical response to unweighing of the running pattern.

Computer methods in biomechanics and biomedical engineering·2013
Same author

Schmallenberg virus: a seroprevalence survey in cattle and sheep, France, winter 2011-2012.

The Veterinary record·2013

Area of Science:

  • Medical Acoustics
  • Signal Processing
  • Machine Learning for Healthcare

Background:

  • Respiratory sound analysis is crucial for diagnosing conditions like asthma and COPD.
  • Accurate classification of normal versus abnormal respiratory sounds (e.g., wheezing) remains a challenge.
  • Traditional methods often struggle with the complexity and variability of lung sound signals.

Purpose of the Study:

  • To develop and evaluate a novel cepstral analysis method using Gaussian Mixture Models (GMM) for respiratory sound classification.
  • To compare the performance of the proposed GMM-based approach against established classifiers like Vector Quantization (VQ) and Multi-Layer Perceptron (MLP) neural networks.
  • To investigate the effectiveness of Mel-Frequency Cepstral Coefficients (MFCC) and subband based Cepstral parameters (SBC) for feature extraction in respiratory sound analysis.

Related Experiment Videos

Main Methods:

  • Respiratory sound signals were segmented into overlapped frames.
  • Reduced-dimension feature vectors were extracted using Mel-Frequency Cepstral Coefficients (MFCC) and subband based Cepstral parameters (SBC).
  • Classification was performed using Gaussian Mixture Models (GMM), with comparisons to Vector Quantization (VQ) and Multi-Layer Perceptron (MLP) classifiers. Post-processing techniques were applied to enhance results.

Main Results:

  • The proposed Gaussian Mixture Model (GMM) method demonstrated effective classification of respiratory sounds into normal and wheezing categories.
  • Feature vectors derived from Mel-Frequency Cepstral Coefficients (MFCC) and subband based Cepstral parameters (SBC) proved valuable for distinguishing sound types.
  • The GMM approach, particularly with post-processing, showed competitive or superior performance compared to VQ and MLP classifiers in preliminary evaluations.

Conclusions:

  • Cepstral analysis combined with Gaussian Mixture Models (GMM) offers a promising approach for automated respiratory sound classification.
  • The feature extraction techniques (MFCC, SBC) are suitable for characterizing relevant information in respiratory sounds for machine learning.
  • Further refinement with post-processing can significantly improve the accuracy of wheezing detection and respiratory sound analysis systems.