AUTHOR=Zhang Fuchun , Li Meng , Song Li , Wu Liang , Wang Baiyang TITLE=Multi-classification method of arrhythmia based on multi-scale residual neural network and multi-channel data fusion JOURNAL=Frontiers in Physiology VOLUME=Volume 14 - 2023 YEAR=2023 URL=https://www.frontiersin.org/journals/physiology/articles/10.3389/fphys.2023.1253907 DOI=10.3389/fphys.2023.1253907 ISSN=1664-042X ABSTRACT=The ECG contains information about key features of arrhythmias, and extracting these implicit features is essential to identify arrhythmias. Therefore, in order to effectively extract the characteristics of ECG data and realize the automatic detection of arrhythmias, a multi-classification method of arrhythmias based on multi-scale residual neural network and multi-channel data fusion is proposed. First, the characteristics of the single-lead ECG signal are extracted and con-verted into two-dimensional images, and the feature datasets are labeled and divided according to different categories of arrhythmia. The improved residual neural network then trains the training set to obtain a classification model of the neural network. Finally, classification models are used for the automatic detection of arrhythmias in motion. The model is evaluated by using the ECG image of multi-lead information fusion, and the accuracy of the classification model is as high as 99.60%, which has high accuracy and generalization ability.