Utilization of Residual CNN-GRU With Attention Mechanism for Classification of 12-lead ECG

Cardiac diseases are the most common cause of death. The fully automated classification of the electrocardiogram (ECG) supports early capturing of heart disorders, and, consequently, may help to get treatment early. Here in this paper, we introduce a deep neural network for human ECG classification into 24 independent groups, for example, atrial fibrillation, 1st degree AV block, Bundle branch blocks, premature contractions, changes in the ST segment, normal sinus rhythm, and others. The network architecture utilizes a convolutional neural network with residual blocks, bidirectional Gated Recurrent Units, and an attention mechanism. The algorithm was trained and validated on the public dataset proposed by the PhysioNet Challenge 2020. The trained algorithm was tested using a hidden test set during the official phase of the challenge and obtained the challenge validation score of 0.659 as entries by the ISIBrno team. The final testing scores were 0.847, 0.195, −0.006, and 0.122, for testing sets I, II, III, and full test set, respectively. We have obtained 30th place out of 41 teams in the official ranking.