Identification of Elevated ST Segment and Deep Q Type MI Variant Using Cross Wavelet Transform and Hierarchical Classification From ECG Signals

Abstract

Coronary artery disease being one of the leading causes of mortality in the world requires prompt attention and for which technological intervention is helpful. The present manuscript proposes a cross wavelet transform based method for classification of Type I Inferior Myocardial Infraction (ST elevated and attenuated T wave) and Type II Inferior Myocardial Infarction (elongated Q wave and inverted T wave morphology). Cross-examination of cross-correlation between two time series reveals localised similarities in time-scale where distinguishing parameters for classification are extracted from wavelet cross spectrum and wavelet coherence. When ECG pattern in the QT region differ from the normal cardiac cycle and a variation is observed in the QT region of lead III, which gives an indication of Inferior MI. The WCS and WCOH of the ECG spectrum together show substantive spectral variations in the QRS complex region (R1) and over the T wave occurrences (R2). The classification of two types of MIs and normal parameters are identified from the corresponding regions R1 and R2. This method gives an accuracy of 99.43% for normal and abnormal class and 88.5 and 87.02% for Type I and Type II Inferior MI respectively.