An algorithm for classification of ambulatory ECG leads according to type of transient ischemic episodes

Abstract

We developed and evaluated an algorithm for classification of ECG leads in ambulatory records according to type of transient ischemic ST segment episodes using the Long- Term ST Database. The algorithm robustly generates ST level function in each ECG lead, tracks non-ischemic ST changes to construct the ST reference function and subtracts it from the ST level function to obtain the ST deviation function. Then the algorithm using statistical moment of the histogram of the ST deviation function given lead classifies the lead according to type of transient ischemic ST episodes (elevations, depressions). The algorithm correctly classified all 9 ECG leads with elevated and 89 out of 90 ECG leads with depressed transient ischemic ST episodes.