부산대학교 도서관

With the development of deep learning, neural networks have been applied to detect myocardial infarction or culprit vessels using electrocardiograms. Although data augmentation is a widely used technique to improve the performance of neural networks, few augmentation methods are designed for the detection of ST-segment elevation myocardial infarction and culprit vessels. The main purpose of this paper is to improve the accuracy in detecting ST-segment elevation myocardial infarction and culprit vessels with a novel ECG data augmentation algorithm. We propose an augmentation scheme, which synthesizes new samples by splicing and transforming heartbeats. To splice heartbeats without manually labeled segmentation points, we design an automatic detection algorithm for a splicing point. We applied our augmentation method to four neural networks to demonstrate its effectiveness. The experimental results show that the largest increases in accuracy and macro-Fl scores of the models are 6.41% and 5.73%, respectively.