Investigation and analysis of ecg signals with advanced signal processing techniques for cardiac arrhythmia classification

Abstract

An electrocardiogram is a non-invasive procedure commonly used to diagnose and monitor patients. It is a record of the electrical activity of the heart. The information collected by an ECG helps understand the functioning of the heart and cardiovascular systems. It can help identify the early signs of heart diseases and provide appropriate treatment. Unfortunately, it is very challenging to analyze an electrocardiogram s long-term records in a short time and also human eye cannot detect subtle changes in the signal. A computer-assisted diagnosis system is needed to monitor the heart s electrical activity. This system can detect abnormal heart rhythms known as cardiac arrhythmia. A cardiac arrhythmia is an irregular heart rhythm disorder that occurs when the heart s electrical activity changes. The considerable variation in the temporal and the morphological characteristics of the data collected during an ECG study is also a significant limitation of the automatic data analysis. This thesis aims to develop advanced signal processing techniques that automatically detect changes in the heart s electrical activity. The main objective of this thesis is to develop a simple and cost-effective method that can be used to detect the cardiac beat after taking an electrocardiogram. This thesis work aims to improve the accuracy of cardiac beat detection by developing a new classification algorithm capable of handling noisy conditions. The proposed algorithm utilizes various classification techniques and features to improve the accuracy of detecting cardiac arrhythmia. In the first work, we proposed a novel technique for effectively newlinediagnosing cardiac arrhythmias from the ECG signal. The proposed method consists of three folds. First, we have found the optimal decomposition level (Jopt) with the optimized quality factor(Q), and the redundancy rate(r) value using the particle swarm optimization (PSO) optimized tunable Q wavelet transform (TQWT) to extract features by decomposing the cardiac arrhythmia ECG signals.