Diagnosing and evaluating the severity of chronic obstructive pulmonary disease based on the time-frequency features of the S transform applied to the lung sound signal

Amin Eskandari, Mahsa; Rashidi, Saeid

doi:10.24200/sci.2023.61685.7440

Diagnosing and evaluating the severity of chronic obstructive pulmonary disease based on the time-frequency features of the S transform applied to the lung sound signal

Document Type : Article

Authors

¹ Faculty of Medical Sciences and Technologies, Science and Research Branch, Islamic Azad University,Tehran, Iran

² Faculty of Medical Sciences and Technologies, Science and Research Branch, Islamic Azad University, Tehran, Iran

10.24200/sci.2023.61685.7440

Abstract

Chronic Obstructive Pulmonary Disease (COPD) is a common respiratory disease characterized by chronic inflammation of the lung airways and destruction of lung tissue that leads to air restriction. Asthma and Chronic Obstructive Pulmonary Disease are the most common respiratory diseases that cause the death of about 180,000 people worldwide every year, and the death rate of COPD is eight times higher than the death rate of asthma. It is the third leading cause of death worldwide. Time-frequency transformation has been used to diagnose and evaluate the severity of this disease using recorded signals, which are dynamic and non-static. In this research, the S transform is used as a tool to extract features from the lung signal. S transform has a higher frequency resolution than the wavelet transforms at low frequencies, and at high frequencies, it has a lower frequency resolution but a higher time resolution. After feature extraction using S transformation, mathematical statistics were applied to reduce feature dimensions. The results indicate that with K-fold validation for KNN classification, the accuracy, precision, and sensitivity values are 98.39%, 97.45%, and 93.88%, respectively, and for SVM, the results are 95.23%, 92.59%, and 83.33%.

Keywords