Electrocardiosignal identification method based on generative adversarial networks and convolution recurrent neural networks

Abstract

The invention provides a single-lead electrocardio abnormal signal identification method based on generative adversarial networks and convolution recurrent neural networks, and mainly solves the problem that data concentration samples are unbalanced. Categories being low in data concentration data volume are subjected to data enhancement, and then identification and classification of electrocardioabnormal signals are performed, so that reference is provided for assistance of doctors, the wrong diagnosis rate and missed diagnosis rate are reduced, and the workload of the doctor can be alleviated. The generative adversarial networks are used for enabling the data concentration samples to achieve the relative balance, so that training the convolution recurrent neural networks can be realized, and good classification effects can be achieved.

Description

technical field [0001] The invention relates to the field of electrocardiographic signal identification and classification, in particular to a single-lead electrocardiographic abnormal signal identification method based on a generative confrontation network and a convolutional cyclic neural network. Background technique [0002] Cardiovascular disease (CVD for short) refers to a series of diseases related to the heart or blood vessels, also known as circulatory system diseases. For the diagnosis of heart disease, electrocardiogram (Electrocardiogram, ECG or EKG) is a transthoracic method that records the electrophysiological activity of the heart in units of time, captures its electrical signal through electrodes placed on the skin, and draws it into a line and records it. Diagnostic technology. As a non-invasive recording method, the electrocardiogram is the most widely used and authoritative. [0003] In recent years, with the improvement of fuzzy recognition, artificial...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to provide a single-lead ECG abnormal signal recognition method based on generative confrontation network and convolutional cyclic neural network, which mainly solves the problem of unbalanced samples in the data set, and classifies the categories with less data in the data set Carry out data enhancement, and then carry out identification and classification of abnormal ECG signals to assist doctors in providing references, reduce misdiagnosis and missed diagnosis rates, and reduce doctors' workload; application of generative confrontation network makes the samples in the data set reach a relative balance, so as to perform convolution Training of recurrent neural network to achieve better classification effect

[0008] Step 2, use the generator of the deep convolution to generate the confrontation network DCGAN to generate ECG data with a small number of samples in the actual data set, and solve the problem of sample imbalance in the actual data set, so that the sample categories with a small amount of data in the actual data set are classified into data Enhanced to get the final training set;

Images