Automatic Diagnosis of Cardiac Disease from Twelve-Lead and Reduced-Lead ECGs Using Multilabel Classification

Variational autoencoder–based neural electrocardiogram synthesis trained by FEM-based heart simulator

“…ECG is a common tool for monitoring cardiac health and detecting heart disease in medical practice. 1 , 2 , 3 , 4 , 5 , 6 , 7 , 8 Various tools and techniques to analyze ECG signals and quickly detect abnormalities have been devised over many decades in order to improve medical care. In recent years, advances in machine learning technologies such as deep neural networks (DNNs) 4 , 9 , 10 have led to rapid development in research that seeks to find useful clinical clues from real ECGs.…”