Deep Learning Algorithm for Classifying Dilated Cardiomyopathy and Hypertrophic Cardiomyopathy in Transport Workers

Abstract

Automatic classification of the different types of cardiomyopathies is desirable, but has been done less with a convolutional neural network (CNN). The aim of this study was to evaluate currently available CNN models for the classification of cine magnetic resonance images (cine-MR) of cardiomyopathies. In this paper, we developed an echo-based wrinkled neural network algorithm for automatic classification of cardiomyopathies. Dilated cardiomyopathy (DCM) is a heart muscle disease with enlargement of the left ventricle or both ventricles and systolic dysfunction. It is an important cause of sudden cardiac death and heart failure and is the most common indication for heart transplantation. MRI of the heart diagnoses severe heart diseases such as myocardial damage and valve problems. This algorithm allows classification of dilated (DCM) cardiomyopathy and hypertrophic cardiomyopathy (HCM) in transport workers with 96.4% accuracy. With the improvement of this approach, it is possible to develop an intelligent system and expert systems to help make decisions aimed at early detection, classification and diagnosis of symptoms of heart disease in transport industry workers.