A NOVEL METHOD FOR DETECTION OF ISCHEMIA IN ELECTROCARDIOGRAM SIGNALS

Abstract

We have proposed a novel method for detection of ischemic events based on morphological features are
extracted from ST-segment abnormalities in electrocardiogram signals named modified isoelectric energy measuring
function (MIEMF). ECG signals are first denoised before going through the delineation process. The ECG data is
first loaded from the European ST-T Database, and then wavelet transforms are used to normalize the ECG to
baseline references and removed the noises. Next, features of the ECG signal are detected, and from these features,
ST-segment is defined and calculated ST deviation for each beat, whether it is normal, elevated, or depressed. The
collected features are subsequently sent to clinicians via an Internet of Things (IoT) cloud channel. After usage of
this approach, we got average sensitivity (SE) is 98.5%, while the average specificity (SP) is 98.3%. These results
outperform those of other methods that have been cited in the literature