Electrocardiogram image processing method and device, medium and electrocardiograph

Abstract

The invention relates to the field of image processing, and discloses an electrocardiogram image processing method and device, a medium and an electrocardiograph. The electrocardiogram image processing method comprises the steps of receiving electrocardiogram images, extracting a feature map of the electrocardiogram image; carrying out dimension reduction processing on the feature map to obtain anattention map; and extracting a feature matrix from the feature map and the attention map by using bilinear attention pool processing, performing adaptive weight learning and weighted fusion on the feature matrix by using multi-head self-attention processing to obtain an expression matrix, and performing multi-label classification on the electrocardiogram image based on the expression matrix. According to the method, electrocardiogram images can be directly interpreted, the method is not limited to use of traditional digital signals, meanwhile, tiny differences in electrocardiogram pieces canbe grasped, and therefore electrocardiogram anomalies can be classified, and the anti-noise capacity is high.

Description

technical field [0001] The present application relates to the field of image processing, in particular to an electrocardiogram image processing method, device, medium and electrocardiograph. Background technique [0002] The electrocardiogram is used to reflect the electrical excitation process of the heart. It is an important clinical method for doctors to perform heart examination and diagnosis. It is generally divided into static electrocardiogram, dynamic electrocardiogram and exercise electrocardiogram. Figure three kind. In the existing technology, the artificial intelligence framework of deep learning is used to learn and process the ECG signal, so as to perform various abnormal types of reasoning, and mainly serve the Xintie product, and the common single-lead ECG based on the Xintie product Analysis of arrhythmia abnormalities. Therefore, there are the following deficiencies in the prior art: on the one hand, signal processing is based on digital signal input, and...

AI Technical Summary

Problems solved by technology

Therefore, there are the following deficiencies in the prior art: on the one hand, signal processing is based on digital signal input, and because of the limitation of electrocardiogram equipment and the storage settings of hospital information systems, a large number of electrocardiograms are stored in the format of pictures in the clinical system, based on digital The signal limits the application in actual scenarios, and most of them are used in the intimate equipment, and cooperate exclusively with the equipment, so the digital signal of the equipment can be obtained directly, which makes the scalability low; on the other hand, most of the existing technologies are based on A single-lead ECG, rather than a clinical-grade 12-lead ECG, allows limited arrhythmia identification

Images