Electrocardiogram waveform data processing method and device and server

Abstract

The invention provides an electrocardiogram waveform data processing method and device and a server. The method herein includes: acquiring electrocardiogram waveform data; preprocessing the electrocardiogram waveform data, inputting the preprocessed electrocardiogram waveform data to a first convolution model, and performing first classification on the electrocardiogram waveform data to obtain first data classification results; inputting the preprocessed electrocardiogram waveform data to a second convolution model to obtain second data classification results; inputting the preprocessed electrocardiogram waveform data to a third convolution model to obtain third data classification results; inputting the preprocessed electrocardiogram waveform data to a fourth convolution model to obtain fourth data classification results; subjecting the four data classification results to weighted averaging to obtain final data classification results. The four convolution models are utilized to classify electrocardiogram waveform data and are suitable for electrocardiogram waveform data; data classification results are more diverse and accurate.

Description

technical field [0001] The invention relates to the technical field of electrocardiographic waveform data analysis, in particular to a processing method, device and server for electrocardiographic waveform data. Background technique [0002] Electrocardiogram is a standard diagnostic tool for non-invasive monitoring of heart electrical activity. In order to detect occasional arrhythmias, patients are often required to monitor ECG for several hours or even longer. The collected ECG waveform signals are very large. The doctor's diagnosis is extremely time-consuming and experience-consuming, so automatic classification of ECG waveform signals in the ECG is a very important research and technology, and it is a more effective auxiliary diagnosis method. [0003] The ECG waveform signal is a one-dimensional signal, and the network model used in the prior art for ECG waveform processing is generally a two-dimensional convolution model and the number of model layers is generally wit...

AI Technical Summary

Problems solved by technology

[0003] The ECG waveform signal is a one-dimensional signal, and the network model used in the prior art for ECG waveform processing is generally a two-dimensional convolution model and the number of model layers is generally within three layers. The adaptability of electrical signals is poor, and the classification results of ECG waveform data obtained through the network model are relatively simple

Images