ECG Diagnostic Classification Based on Neural Networks

Abstract

Heart arrhythmias result from any disturbance in the rate, regularity, and site of origin or conduction of the cardiac electric impulse. Premature ventricular contraction (PVC), left bundle brunch block (LBBB), and right bundle branch block (RBBB) are the three cardiac arrhythmias which can lead to or indicate the risk of heart failure. Automated analysis of the digital 12-lead ECG involves signal analysis and diagnostic classification. We will design a pattern classifier based on neural network (NN) able to recognize the above cardiac arrhythmias. The classifier will be tested on MIT-BIH arrhythmia database in order to detect PVC, LBBB, RBBB arrhythmias and normal heart rhythm. A major purpose of this research is to further provide clinicians with insight into the generally missing link between technology and its consequences for clinical ECG interpretation. Today, various digital-signal-processing methods are applied to the ECG to identify, extract and analyze the different ECG signal components. In this large set of signal-processing tools we propose to use wavelet transformation in order to describe time and frequency characteristics of ECG waves.