Detection of heart sounds S1 and S2 using optimized S-transform and back — Propagation Algorithm

Abstract

This paper describes a method for the detection of heart sounds S1 and S2 using Back-Propagation Algorithm (BPA). The proposed method works satisfactorily on the heart sounds with murmur or noisy data. The work can be divided into four main parts: (i) localization by optimized S-transform, (ii) boundary detection of heart sounds S1 and S2 based on Shannon energy of S-transform, (iii) feature extraction using Singular Value Decomposition (SVD), and (iv) classification using ANN tool known as BPA. In this paper, we have compared the performance parameters like sensitivity and specificity of BPA results with other existing methods and improvement is observed in their values.