Classification of respiratory diseases based on lung and cough sound using convolutional neural network

Abstract

Accurate non-invasive methods for detecting respiratory diseases, including COVID-19, are needed to suppress the potential of infection deployment to medical personnel when assessing the patient. In this report, we proposed respiratory diseases detection based on the cough and lung sounds of the patient. The recorded cough and lung sounds are processed for obtaining the mel frequency cepstral coefficient (MFCC) which is used as classification features. The features are then used for data training on deep learning models using convolutional neural networks (CNN). The result showed that the model can classify respiratory diseases with high accuracy compared with the prior studies.