A novel method for wet/dry cough classification in pediatric population

Abstract

Cough is one of the early symptoms of the respiratory tract infections. Cough can be classified into wet and dry classes based on the mucus. In pediatric population, knowing the type of cough is useful for diagnosis as well as treatment. Currently, the classification of coughs is carried manually during the physical examination. Manual classification is tedious and subjective. The results depend heavily on the experiences and the skills of the physicians. In this paper, we propose a novel feature set called intra-cough shimmer jump (ICSJ) and intra-cough Teager energy operator jump (ICTeJ) to discriminate dry/wet coughs. The proposed method achieved 76% of accuracy, 77% of sensitivity and 75% of sensitivity. The Kappa agreement was 0.52, indicated a moderate agreement with the classification from the pediatricians. Our method can be developed into a non-invasive tool for assessing cough in the resource limited-settings.