Wavelet Packet VS Backpropagation for localization and Classification PCG Signals

• Published in: IOP conference series. Materials Science and Engineering 2020-07, Vol.881 (1), p.12104
• Main Authors: Abed, Wasam Samer, Fayadh, Rashid Ali
• Format: Article
• Language: English
• Subjects: Backpropagation; Cardiac Auscultation; Decomposed; Phonocardiography; Reconstruction; WPT
• ISSN: 1757-8981; 1757-899X
• DOI: 10.1088/1757-899X/881/1/012104

Phonocardiography signals (PCG) show the enrolment of sounds and distortions generated from cardiac auscultation. Cardiac signal analysis is crucial for the diagnosis of various diseases. In previous years, there are several groups and different techniques, freshness, and methods have been proposed to analyze the cardiac signal. Auscultation is a manner in which a stethoscope is used to hearken to the cardiac sound. Structural trouble of the heart is often reflected in the cardiac sound created and Listening to the heart's sound helped doctors diagnose and predict diseases. Whilst a cardiac sound examine via Listening is appropriate as a scientific tool, it is tough to analysis PCG signals in the time(T) or frequency(F) scale. The PCG signals have many benefits over conventional Listening, in that they may be rebuilt and analyzed for time(T) and frequency(F) information. Using a wavelet packet transform(WPT). Where the signal is decomposed and rebuild without the first-rate loss of data within the signal content. Reconstruction mistakes can be thoughtfulness an important piece of information in classifying the pathological severity of phonocardiography signals. In this paper we will focus on how to choose the level and the mother wave of the wave to cry out so that it is appropriate to analyze the cardiac signal in good mathematical and analytical ways to train the neural network (error backpropagation) on it. It will be explained in the following details.