Prony residual analysis for the identification of cardiac arrhythmias

Abstract
An investigation into the use of Prony modeling for the identification of late potentials (LPs) from signal-averaged electrocardiograms (SAECGs) is described. We develop a noninvasive method, based on the short-time Prony modeling algorithm, to extract a diagnostic feature, dubbed the Prony residual marker (PRM), from the SAECG waveform. The PRM is used as a marker for the presence of LPs and is thus used to predict the diagnostic outcome of invasive electrophysiologic study.