Electrocardiogram data classification method, device and system

Abstract

The present invention provides a method, device and system for classifying electrocardiogram data, wherein the method includes: using the echo state network to convert the electrocardiogram data to be classified into electrocardiogram model data; inputting the electrocardiogram model data to a pre-trained normal classifier, and obtaining the first A classification result; input the electrocardiogram model data to one or more pre-trained abnormal classifiers, and obtain one or more second classification results; for the second classification result output by each abnormal classifier: if the first If the classification result indicates that the ECG model data is normal and the second classification result indicates that the ECG model data is not abnormal, then it is determined that the ECG data to be classified is normal; if the first classification result indicates that the ECG model data is abnormal and the second If the result of the binary classification indicates that the data of the electrocardiogram model is abnormal, it is determined that the data of the electrocardiogram to be classified is abnormal. This application can accurately judge the electrocardiogram data and improve the accuracy.

Description

technical field [0001] The present application relates to the field of communication technology, in particular to a method, device and system for classifying electrocardiogram data. Background technique [0002] Electrocardiogram is an important diagnostic tool for many cardiovascular diseases. Now, the use of artificial intelligence to assist the diagnosis of electrocardiogram is one of the important research fields of today's medicine. Existing intelligent classification technologies for ECG mainly focus on traditional machine learning and deep machine learning. [0003] Traditional machine learning will perform feature extraction on ECG for classification. When there are some uncertainties in ECG, accurate feature extraction cannot be performed. Therefore, the classifier obtained by traditional machine learning can not accurately judge ECG data. [0004] Compared with traditional machine learning, deep machine learning has fewer limitations, but deep learning requires a ...

AI Technical Summary

Problems solved by technology

[0003] Traditional machine learning will perform feature extraction on the ECG for classification. When there is some uncertainty in the ECG, accurate feature extraction cannot be performed. Therefore, the classifier obtained by using traditional machine learning is inaccurate in determining the ECG data.

[0004] Compared with traditional machine learning, deep machine learning has fewer limitations, but deep learning requires a large amount of training data as support, but for a certain disease, the ECG data used as training samples may not reach the amount of deep machine learning, so use The classifier obtained by deep machine learning is inaccurate in the judgment of ECG data

Images