Ischemic and non-ischemic heartbeat classifier for portable automatic detection systems

This article presents a new heartbeat classifieroriented to an automatic cardiac ischemia detection system aimedat the medical control of the population with cardiac disorders (ahigher risk group of severe COVID-19). The algorithm focuses ondistinguishing ischemic heartbeats from non-ischemic based onmethods and techniques proposed by medical experts in cardiacischemia diagnoses. The proposed algorithm was tested with 80 electrocardiogram recordings obtained from the Long-term STdatabase of Physionet with a 100,112 heartbeat population. Theheartbeat population contains different types of non-ischemiccases. The algorithm exhibits a sensitivity of 98.7 percent andspecificity of 97.23 percent. The digital synthesis, in a 180nmCMOS process with a dynamic power consumption of 13.15 microWatt demanding an area of 490x400 micrometers, attests to thealgorithm benefits.