Exploring The Potential of Using Discrete Wavelet Transform (DWT) for Heart Sound Analysis

Abstract

This paper delves into the effectiveness of the Discrete Wavelet Transform (DWT) in analyzing cardiac sounds, highlighting its benefits, practical applications, and associated challenges It reviews various DWT-based techniques for preprocessing, feature extraction, and classification of cardiac sounds, emphasizing their success in detecting anomalies and enhancing diagnostic precision The study showcases DWT's potential as a powerful tool for improving cardiac sound analysis
The proposed method was evaluated, achieving an accuracy of 84%, with sensitivity ranging from 63% and specificity from 62% These results underscore the method's reliability in contributing to neural network systems for classifying cardiac sound signals as normal or abnormal The Adaptive Neuro Fuzzy Inference System (ANFIS), when combined with DWT attributes, emerges as an effective tool, showcasing its capability in the PhysioNet Challenge 2016.