Combined weighted feature extraction and deep learning approach for chronic obstructive pulmonary disease classification using electromyography

The COVID-19 outbreak has led to a rise in respiratory disease-related deaths, including Chronic Obstructive Pulmonary Disease (COPD). Early diagnosis of COPD is crucial, but it can be challenging to distinguish between different chronic pulmonary diseases due to their similar symptoms, leading to misdiagnosis and time-consuming manual inspections. To address this issue, this paper explores the use of a deep learning model to differentiate COPD from other lung diseases using lung sound captured during Electromyography (EMG). The model includes steps such as noise removal, data augmentation, combined weighted feature extraction, and learning. The model's efficacy was evaluated using various metrics, including accuracy, precision, recall, F1-score, kappa coefficient, and Matthew’s correlation coefficient (MCC), with and without augmentation. The results show that the model achieved 93% accuracy and outperformed other existing state-of-the-art deep learning models, increasing the robustness of clinical decision-making.