QT-STNet: A Spatial and Temporal Network Combined with QT Segment for MI Detection and Location

Myocardial infarction (MI) is a cardiovascular disease with high mortality, which can be diagnosed by electrocardiogram (ECG). Therefore, this paper proposes a spatial and temporal network combined with QT segment (QT-STNet) to detect and locate MI. First, QT segment is extracted by QT segment extraction module. Next, the QT segment is concatenated with the original signal to enhance the importance of the key information. Then, considering the spatial and temporal features of ECG, DenseNet is used to extract spatial features and Transformer is used to extract temporal features. Finally, the proposed method is test on PTB-XL dataset, and the precision, recall, F1 score and hamming loss are 0.881, 0.881, 0.876 and 0.051 respectively. The results show that the method is superior to other methods.