Date of Award

3-2-2021

Publication Type

Master Thesis

Degree Name

M.A.Sc.

Department

Electrical and Computer Engineering

Keywords

Adaptive multilevel thresholding, Discrete wavelet transform, QRS complex detection, Segmentation, Statistical false peak elimination

Supervisor

Esam Abdel-Raheem

Rights

info:eu-repo/semantics/embargoedAccess

Abstract

Heart beats are important aspects of the study of heart diseases in medical science as they provide vital information on heart disorders and diseases or abnormalities in the heart rhythm. Each heart beat provides a QRS complex in the electrocardiogram (ECG) which is centered at the R-peak. The analysis of ECG is hindered by low-frequency noises, high-frequency noise, interference from P and T waves, and change in QRS morphology. Therefore, it is a major challenge to detect the QRS complexes using automatic detection algorithms.This thesis aims to present three new peak detection algorithms based on a statistical analysis of the ECG signal. In the first algorithm, a novel method of segmentation and statistical false peak elimination is proposed. The second algorithm uses different levels of adaptive thresholds to detect true peaks while the third algorithm combines and modifies the two proposed algorithms to provide better efficiency and accuracy in QRS complex detection. The proposed algorithms are tested on the MIT-BIH arrhythmia and provides better detection accuracy in comparison to several state-of-the-art methods in the field. To evaluate the performance of the proposed method, the merits of evaluation consider the number of false positives and negatives. A false positive (FP) is the result of a noise peak being detected and a false negative (FN) occurs when a beat is not detected at all. The methods emphasize better detection algorithms that detect peaks efficiently and automatically without eliminating the high-frequency noise completely and hence reduces the overall computational time.

Share

COinS