Cough Based Lung Infection Detection Using Deep Learning

Abstract

Pulmonary infections are a major global health concern that can result in a significant number of fatalities annually, regardless of age. Early diagnosis of lung disorders is essential to enable medical professionals to treat the afflicted individual. Many deep learning algorithms have been utilized in current systems to identify lung illnesses from cough sounds, but the results have not proven very accurate. This work presents a novel deep learning algorithm-based method for the early identification of lung infections using cough sounds. The suggested system uses the deep learning algorithm, RESNET-18, to analyze and categorize cough sounds. By listening to a person's cough sounds, it is possible to determine whether or not they may have a lung disease, such as pneumonia, pulmonary edema, asthma, TB, COVID19, pertussis, or another respiratory illness. It offers a safe and economical way to determine the likelihood of lung infections.