Myocardial Infarction Localization and Blocked Coronary Artery Identification Using a Deep Learning Method

摘要

Myocardial Infarction (MI) is a fatal heart attack that causes millions of death in the world every year and the mortality of this disease is extremely high. Therefore, the precise diagnosis and the instant alert of MI is vital for the patients in hospital. In this paper, we have provided a deep learning method using Lenet to determine whether there is MI in the patient and the necrosis region of MI, from which the model can also provide the prediction of the blocked coronary artery. The accuracy of the model using single lead of ECG independently can achieve more than 99% and even the precision, sensitivity and specificity can exceed more than 99%, which outperforms most of the algorithms.