Exploring Cough Sounds in Chronic Obstructive Pulmonary Disease (COPD) Patients Using Machine Learning Techniques

We classified respiratory sounds that indicate the early stages of coughing in patients with chronic obstructive pulmonary disease (COPD). Deep learning techniques with a voice spectrogram were utilized for respiratory sound detection. Convolutional neural networks (CNNs) were applied to classify COPD by extracting features from the respiratory sounds. The optimized model achieved an accuracy rate of 91 %. Since the pretraining database was sourced from publicly available data from the Internet, the model was enhanced by incorporating COPD sounds provided by doctors as feedback.