Classification of respiratory pathology in pulmonary acoustic signals using parametric features and artificial neural network

Pulmonary acoustic signal analysis provides essential information on the present state of the Lungs. In this paper, we intend to distinguish between normal, airway obstruction pathology and interstitial lung disease using pulmonary acoustic signal recordings. The proposed method extracts Mel frequency cepstral coefficients (MFCC) and AR Coefficients as features from pulmonary acoustic signals. The extracted features are then classified using Artificial Neural Network (ANN) classifier. The classifier performance is analysed by using confusion matrix technique. A mean classification accuracy of 92.59% and 91.69% was reported for the MFCC features and AR coefficients features respectively. The performance analysis of the ANN classifier using confusion matrix revealed that normal, airway obstruction and interstitial lung disease are classified at 92.75%, 91.30% and 92.75% classification accuracy respectively for the MFCC features. Similarly, normal, airway obstruction and interstitial lung disease are classified at 92.75%, 91.30% and 89.85% classification accuracy respectively for the AR coefficient features. The analysis reveals that the proposed method shows promising outcome in distinguishing between the normal, airway obstruction and interstitial lung disease.