Accuracy assessment of sensed biomedical images for myocardial infarction prediction

Abstract

Myocardial infarction (MI) can be defined from a number of different perspectives related to clinical, electrocardiographic (ECG), biochemical and pathologic characteristics. The term MI also has social and psychological implications, both as an indicator of a major health problem and as a measure of disease prevalence in population statistics and outcomes of clinical trials. In the distant past, a general consensus existed for the clinical entity designated as MI. In studies of disease prevalence by the World Health Organization (WHO), MI was defined by a combination of two of three characteristics: typical symptoms (i.e., chest discomfort), enzyme rise and a typical ECG pattern involving the development of Q waves. Biomedical sensors dedicated to acquire signals from cardiac instrumentation, even if sophisticated, cannot precisely reveal and help doctors to understand, at a glance, pathologies leading towards MI. This paper traces out an integrated algorithm based on a combination of level set evolution and variational approach according to Mumford-Shah model.