Method for detecting key area of electrocardiosignal based on multiple views

Abstract

The invention discloses a method for detecting a key area of an electrocardiosignal based on multiple views, belongs to the technical field of medical technology and deep learning, and solves the problems that in the prior art, the visualization effect is rough, and accurate positioning cannot be achieved. The method sequentially comprises the following steps of 1) collecting the electrocardiosignal; 2) extracting a shallow view of the electrocardiosignal; 3) constructing a convolutional neural network; 4) acquiring a prediction result Y of the electrocardiosignal based on the convolutional neural network; 5) training the convolutional neural network based on the prediction result Y; 6) obtaining a saliency map of the electrocardiosignal to be recognized based on the prediction result Ys and the Grad-CAM technology; and 7) visually displaying an output picture of a key focus area of the convolutional neural network. The method is used for detecting the key area of the electrocardiosignal.

Description

Technical field [0001] A method for multi-view ECG focus detection area based on the ECG signal for detecting the focus area, a medical technology and technical field depth study. Background technique [0002] Arrhythmia middle finger in the medical field is too fast or too slow heart rate, and heart rate than the general range. This phenomenon is usually due to cardiac causes abnormal automaticity or conduction disorders tachycardia, bradycardia or arrhythmia. It causes an irregular heartbeat often associated with unhealthy living habits, such as mental stress, heavy smoking, drinking, drinking tea or coffee, fatigue, severe insomnia. Therefore, early detection, early prevention, can effectively prevent the occurrence of malignant heart disease. [0003] The main types of cardiac arrhythmia comprise: sinus bradycardia, sinus tachycardia, sinus arrhythmia, extrasystoles, early shrinkage, fibrillation, atrial flutter, paroxysmal ventricular tachycardia, ventricular rhythm tachycar...

AI Technical Summary

Problems solved by technology

[0009] In view of the problems of the above research, the purpose of the present invention is to provide a method for detecting key areas of ECG signals based on multi-view, which solves the problem that the prior art has rough visualization effects and cannot precisely locate key areas

Images