Spatial shape variability analysis of T-waves in the ECG signal

Abstract

This work aims to study the shape variability of the T-waves using multi channel records. The classification between a group of subjects with Myocardial Infarction (MI) and a group of healthy subjects is obtained using a synthetic signal called Integral Shape Averaging (ISA). Our methodology is based on calculation of projection coefficients corresponding to each electrode for both groups of subjects. These coefficients represent a distance between each subject and one of the groups. This latter allows us to compare each subject to each population. These coefficients are obtained by a surface difference between the signal of each electrode in each column and the associated ISA signal. In addition, we seek for the electrode that represents the best separation between the two populations. In this study we improve previous studies and make a comparison with another approach which is a combination between the ISA technique and the calculation of shape difference.