A parallel approach to ECG processing

Abstract

The algorithm presented is based on parallel processing of the Electrocardiogram (:ECG) using the information contained in all 12 -simultaneously recorded-leads. A non-syntactic algorithm is used initially to extract the results in every one lead. The computing for each lead is done in 12 parallel processors and concurrently a cross-correlation and evaluation of the individual results is done in the main processor, in order to make the final decision about each detected characteristic. It turns out that this approach gives better and more reliable information than the known methods which are based on conventional methods.

A Decentralized Multichannel Length Transformation Algorithm and Its Parallel Implementation for Real-Time ECG Monitoring

“…Although we initially followed this method (24), it soon became clear that this approach has some drawbacks. For example, it is quite difficult to compute these factors and the threshold ⌰ for all normal and abnormal beats which can be found in an ECG.…”

Εφαρμογή Παράλληλων Μεθόδων Και Ειδικών Διατάξεων V.L.S.I. Στην Αυτόματη Αναγνώριση Ηλεκτροκαρδιογραφήματος

“…Στο στάδιο αυτό υπάρχουν πάλι δυο προσεγγίσεις. Η πρώτη, την οποία ακολουθήσαμε αρχικά [KOUL96], έγκειται στο να κανονικοποιήσουμε τις μετρήσεις αυτές δίνοντας κάποια βάρη. Η προσέγγιση αυτή είναι μια πολυκαναλική επέκταση του αλγορίθμου που πρότειναν οι Bond et al [BOND93].…”