Automatic Multi-label Classification in 12-Lead ECGs Using Neural Networks and Characteristic Points

Electrocardiogram (ECG) signals are widely used in the medical diagnosis of heart disease. Automatic extraction of relevant and reliable information from ECG signals is a tough challenge for computer systems. This study proposes a novel 12-lead electrocardiogram (ECG) multi-label classification algorithm using a combination of Neural Network (NN) and the characteristic points. The proposed model is an end-to-end model. CNN extracts the morphological features of each ECG. Then the features of all the beats are considered in the context via BiRNN. The proposed method was evaluated on the dataset offered by The First China Intelligent Competition, and results were measured using the macro F1 score of all nine classes. Our proposed method obtained a macro F1 score of 0.878, which is excellent among the competitors.