Identification of Myocardial Ischemic and Infarction Episodes Based on ST Level and Beat Type Re-attribution Method

Abstract

The objective of this study is to establish an efficient and effective recognition system for myocardial ischemic and myocardial infarction episodes in ECG. We first applied a preprocessing algorithm to reduce noise and baseline wander. Then, we simplified the procedures of identifying the important points and defined these points based only on heart rate and the R peak which is relatively unaffected by noise. Thirdly, an ST-deviations-based algorithm was used to identify both myocardial ischemic (MIs) and myocardial infarction (MIn) beats. Finally, a merging algorithm followed by correcting windowing was employed to re-evaluate the attribute of each beat for more accurately identify the beginning and end points of the episodes. The results show that, the proposed method raises the recognition rates from 87.53%, 85.12%, and 80.41%, in identifying MIs, MIn, and normal beats, respectively, to 94.63%, 91.56%, and 92.89%, respectively. The results demonstrate the efficiency and effectiveness of the proposed method in accurately identifying myocardial ischemic and infarction episodes.