Computer-aided electrocardiogram analysis

Abstract

An electrocardiogram analysis package is developed in this thesis. A parabolic smoothing with pre-set threshold for attenuating effects on high frequency components is applied. The GF and operational transforms are studied and examined for their constraints, applications and flexibility. These transforms are modified and adapted for recognition of QRS onset and offset in the Frank Orthogonal Lead System. An algorithm, based on a set of pre-set criteria and thresholds, is developed. Various clinical parameters, vectors and files for analysis. such as segment or wave intervals, magnitude, areas etc. are generated together with data two-dimensional and three-dimensional graphic In addition. a general purpose graphic package is supplied to create the vector displays for ECG analysis. The algorithm has been tested on three hundred cardiac cycles taken from ten patients with both normal and abnormal ECG characteristics.