Phonocardiogram based Method for the Classification of Coronary Artery Diseases

摘要

Cardiovascular diseases are on the top list and affecting many people around the world. In this paper, a novel method is proposed in which signal processing of PCG (Phonocardiography) is used for classifying coronary artery diseases (CAD). The algorithm used an innovative idea in which we have not done segmentation which has decreased computational time, extracted the combination of features (MFCC and Local ternary pattern method), and de-noised the signal by using Empirical Mode decomposition (EMD). We employed (KNN) K-Nearest Neighbor classifier subspace fine for the CAD classification and the performance of the obtained algorithm has been evaluated by using 605 signals of patients with different CAD diseases. The proposed algorithm has achieved 90.1% accuracy.