Detection and Recognition of Cough Sounds Using Deep Learning for Medical Monitoring

Coughing is the most prevalent indication of respiratory illness. The importance of cough detection and classification as a preventive measure against infectious diseases cannot be overemphasized. The importance of cough testing is also relevant to public health. However, accurate classification of cough in noisy environments remains a major challenge. Pattern recognition algorithms can be adapted to work in noisy environments and real-time situations. Our cough detection and classification method is based on the analysis of the acoustic signature of coughs using artificial intelligence and deep learning. Before training the deep neural network model, the MFCC (Mel-Frequency Cepstral Coefficients) method is used to extract the acoustic characteristics of the sounds used. Our system uses a sensor attached to the neck to detect movements of the throat and trachea during coughing. The results of our study on the acoustic analysis of coughing were satisfactory. This model classifies coughs into three classes with an accuracy of 78.09%. We are developing state-of-the-art systems that can effectively detect and classify coughs, even in noisy environments. This means that hospitals, airplanes and noisy public places can use our technology to identify patients in need of medical attention or with symptoms of respiratory disease.