To achieve a rapid assisted diagnosis of the site of myocardial infarction(MI) occurrence by clinical medical devices,a high-accuracy MI site localization algorithm is designed based on the lightweight convolutional neural network of MobileNetV2 combined with the coordinated attention(CA)mechanism.The 12-lead electrocardiogram(ECG)samples of normal and MI cases are filtered from the PTB dataset,and the ECG signals are denoised.The R-peaks of ECG signals are detected by using the differential thresholding method,the heartbeat samples are segmented according to the R-peaks,and the heartbeat data are used to train and test the model designed.The classification performance of the model is evaluated by using accuracy,precision,sensitivity,specificity and confusion matrix.After iter-ating the training set for 60 rounds,the accuracy of the test set reached 99.91％.The results show that the MobileNetV2 model incorpo-rating the CA module is effective for localization of MI sites and helps medical devices to achieve rapid assisted diagnosis of MI.
维普
万方数据