A New Approach to Classifying Myocardial Infarction and Cardiomyopathy Using Deep Learning
Annotatsiya
In this study, for the first time, an automatic diagnostic algorithm for myocardial infarction (MI) and cardiomyopathy was developed based on ECG data obtained at different periods of time. ECG data were used taken from the ECG-VIEW II database. ECG data classification was done by the developed Convolution Neural Network (CNN) model. In order to determine the effectiveness of the proposed model, diseases with similar ECG manifestations (myocardial infarction and cardiomyopathy) were selected, which are still the most misdiagnosed among physicians today. Despite the less of data and dedicated symptoms and the similarity of symptoms of the disease on the ECG, the accuracy of the network test result became high, 91,1%. The conclusion is that with a larger database and more features, or with 12 channel ECG database it is possible to further increase the accuracy of the network, as well as increase capability of the network to differentiate other diseases. Moreover, most importantly, this model has the ability to increase the diagnostic accuracy of 1-channel smartphone ECG devices.