Left ventricular hypertrophy detection method and system based on deep learning early stop mechanism

Abstract

The invention belongs to the field of medical treatment, and particularly provides a left ventricular hypertrophy detection method and system based on a deep learning early stop mechanism, and the method comprises the steps: S1, collecting 12-lead electrocardiosignals of a patient in a resting state to obtain a training set; s2, each electrocardiosignal is judged, if the electrocardiosignal is not left ventricular hypertrophy, the electrocardio tag vector is [0], otherwise, the electrocardio tag vector is [1], and after each electrocardiosignal corresponds to one electrocardio tag vector, a tag set is formed; s3, training the label set by using a convolutional neural network model to obtain a training model; s4, 12 lead electrocardiosignals are sequentially input into the trained training model, and the training model predicts an electrocardiograph tag vector according to the electrocardiosignals; if the output is greater than 0.5, judging that the electrocardiosignal is left ventricular hypertrophy; and if the output is not greater than 0.5, judging that the electrocardiosignal is not left ventricular hypertrophy. According to the scheme, under the condition that the accuracy is not influenced as far as possible, the robustness of the convolutional neural network for electrocardiogram classification is improved.

Description

technical field [0001] The present invention relates to the medical field, and more specifically, to a method and system for detecting left ventricular hypertrophy based on a deep learning early stop mechanism. Background technique [0002] The left ventricle is one of the important components of the heart. Its main function in the body is to receive oxygenated blood and pump it to the aorta to complete the systemic blood supply. With the prolongation of the course of hypertension, etc., the left ventricular muscle wall will gradually thicken, and the thickness or volume of the left ventricular muscle wall exceeds the normal range, which may not only aggravate myocardial ischemia, but also directly lead to death of the patient in severe cases. In the past, because left ventricular hypertrophy was not easy to detect, it often required long-term monitoring and manual ECG monitoring to achieve it, which was time-consuming and labor-intensive. Therefore, it is particularly impo...

AI Technical Summary

Problems solved by technology

[0004] The present invention aims at the technical problem that the robustness of the existing neural network used in the prior art is not strong enough, so that it has the disadvantage of poor classification effect on the test samples, especially the test samples when the data distribution of the training samples is different from that of the training samples.

Images