Intelligent System for Classification of Pulmonary Diseases from Lung Sound

Abstract

Lung disease belongs to the class of fatal disease according to World Health Organization statistics. Asthma and bronchitis are most prominent among these abnormalities. Identification of Lung disease from lung sound analysis is still question mark in medicine. In this paper, Asthma and Bronchitis are identified from Lung sound analysis from application of signal processing techniques. Data is acquired from 50 Asthma, 50 Bronchitis and 50 Normal subjects. Empirical mode decomposition method is used at preprocessing stage. Standard deviation, Shannon energy, peak to peak and root mean square features are estimated and system accuracy on different k-NN classifier is analyzed via Matlab 2019a. The system evidenced greater than 99.30% accuracy. Further improvement can be done in exploring new features of Lung sounds for better and robust classification of different Lung diseases.