Intelligent classification and recognition method of heart beat of electrocardiogram

Abstract

The invention belongs to the technical field of electrocardiosignal recognition, and relates to an intelligent classification and recognition method of heart beat of an electrocardiogram. The intelligent classification and recognition method of heart beat of an electrocardiogram comprises the following steps of obtaining heart beat signal sample data of the electrocardiogram, wherein the sample data comprises heart beat signal data of an electrocardiogram of a healthy person and heart beat signal data of electrocardiograms of different kinds of patients, and preprocessing the sample data; constructing a stacked auto-encoder, and automatically extracting characteristic information in the heart beat signals in an unsupervised manner; based on the characteristic information, establishing a classification and recognition database, training a general-purpose background model, studying general characteristics, and initializing a special recognition model for heart beat through parameters; collecting the heart beat of each patient of first m minutes, and training the special recognition model so as to study heart beat specificity; and testing the recognition accuracy of the model throughsubsequent minute data. According to the intelligent classification and recognition method of heart beat of an electrocardiogram disclosed by the invention, electrocardio recognition accuracy is promoted, real-time continuous monitoring of electrocardiosignals is realized, and the intelligent classification and recognition method of heart beat of an electrocardiogram can be suitable for multiple fields of clinical monitoring, wearable equipment monitoring and the like so as to assist doctors to quickly perform relevant diagnosis.

Description

technical field [0001] The invention belongs to the technical field of electrocardiographic signal recognition, and in particular relates to an intelligent classification and recognition method for electrocardiogram and cardiac beats. Background technique [0002] In recent years, cardiovascular disease accounts for the highest proportion of major disease mortality among Chinese residents. It has become the main cause of death in my country and even in the world, and poses a serious threat to people's health. Common cardiovascular diseases include heart attack, ischemic stroke, hemorrhagic stroke, heart failure, different types of arrhythmias and heart valve disorders. The electrocardiogram signal is a record of heart activity and can be captured by an electrocardiograph. The heart abnormalities caused by the above-mentioned diseases will all appear as some abnormal waveforms in the patient's electrocardiogram signal. The diagnosis of electrocardiogram is often done directl...

AI Technical Summary

Problems solved by technology

The heartbeat waveforms of various heart disease patients are not only abnormal and diverse, but also the real-time recording equipment of the heartbeat signal is often affected by the working environment and introduces a large amount of different types of noise information, making the effective diagnosis of the electrocardiogram very difficult. Challenging Research Questions

The existing ECG automatic diagnosis is roughly divided into three stages: signal preprocessing, feature generation and recognition; although feature extraction has made significant progress, the recognition performance of cardiac beats with serious noise interference is still very low, and it is difficult to meet the requirements of clinical diagnosis; , the recognition classification mainly focuses on the recognition of various types of heartbeats, but does not pay attention to the heartbeat recognition of specific patients under real-time recording conditions, ignoring the differences in heartbeat waveforms between different patients, and the performance of heartbeat recognition for new patients is not ideal. To be further improved

Images