Transfer Learning Aided Classification of Lung Sounds-Wheezes and Crackles

Nowadays, Lung diseases are the most life-threatening disease. In maximum belongings of lung disease are noticed once the illness is in the progressive stage, detection of lung disease can be supportive in to cure in early stages. Nowadays, advancement in technology plays an important role in healthcare. By using electronic stethoscopes, the lung sounds of the patients are recorded. Lung sounds carry important information related to lung diagnosis. The need for identifying lung disease using lung sounds is an active research area in the field of the healthcare domain. Transfer learning plays an important role in the medical system. Here, research paper represents various Machine Learning and Transfer learning approaches for the lung sounds classification. Further, Transfer learning model, the classification will be done on the basis of the RESNET-50 deep network and Mel spectrogram of lung sound signals. These classification models achieve 80% accuracy in lung sounds-wheezes and crackles classification which can be use lung disease diagnosis for future research.