Get started
Work

Classification of Adventitious Lung Sounds: Wheeze, Crackle using Machine Learning Techniques

Year: 2023
Type: article
Abstract: In the past decade, there has been a significant surge of research interest in automated detection of respiratory disorders using advanced stethoscope technology in both academic and industrial domains. Researchers and practitioners aim to achieve more objective respiratory disease diagnosis by harnessing machine learning models. Pulmonologists frequently observe wheeze and crackle in lung sounds during patient auscultation, and machine learning algorithms have been adopted to mitigate the subjective nature of diagnosis. The project aimed to classify four target classes, namely "none, crackle, wheeze, both," utilizing machine learning techniques, specifically Support Vector Machine (SVM) and Decision Tree (DT) classifiers. The primary goal was to train and test these models to accurately identify respiratory conditions based on temporal lung sound signals. Lung sounds, as temporal signals, are generated during the inhalation and exhalation process. Irregular lung sounds can indicate various respiratory conditions, including inflammation, fluid accumulation, or other abnormalities in the airways or lung tissue. Upon training and testing the SVM and DT classifiers on the dataset, the SVM classifier exhibited superior accuracy compared to the DT classifier. Across all performance metrics, the SVM classifier outperformed the DT classifier, establishing it as the most effective choice for classifying the four target classes. The combination of advanced stethoscope technology and machine learning techniques, particularly SVM classification, presents promising results for automated respiratory disorder detection. This approach has the potential to significantly enhance diagnostic accuracy and objectivity in the field of respiratory medicine, benefiting both patients and healthcare professionals. Further research and development in this area hold promise for substantial advancements in respiratory disease diagnosis and treatment.
Language: English
Field: Medicine
Open Access status: diamond
OpenAlex ID: W7217036784
FWCI:
Cites:
Cited by:
Related to: