Voluntary Cough Detection By Internal Sound Analysis

Abstract

Cough can be defined as a forced expulsive onrush, normally against a closed glottis, producing a characteristic three-phase sound, and as a symptom, it can be an indicator of many respiratory diseases. An objective measure of cough would be of use in clinical practice, clinical research and the assessment of novel therapies and pharmaceuticals. In the present work, a method to automatically identify, count and (partly) qualify cough sounds, based on internal sound signals, is proposed. This approach relies on explosive phase detection, because of its acoustic and spectral distinctive characteristics, and its potential for accurate onset detection of cough sounds. The features analyzed, related with tonality, pitch, timbre and frequency, prove to be very relevant in our explosive phase detection approach. Our results show an accurate detection, for a wide testing population with and without respiratory perturbations, which demonstrates the ruggedness of this approach. The internal sound analysis reveals advantageous in external noise reduction, therefore internal sounds are highlighted and better characterized. The explosive phase detection approach demonstrates usefulness in detecting the onset of cough sounds.