Unlock instant, AI-driven research and patent intelligence for your innovation.

Electrocardio waveform positioning and classification model device adopting multi-scale visual field deep learning

A model device and depth of field technology, applied in applications, medical science, sensors, etc., can solve problems such as complex steps, high incidence of electrocardiogram positioning errors, and poor robustness

Pending Publication Date: 2021-03-30
SHANGHAI SID MEDICAL CO LTD
View PDF12 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Then use the pattern recognition method to identify the R wave, and then judge the type of waveform. The diagnostic conclusion cannot be directly mapped to the waveform location, and the steps are complicated.
Second, because the heart rate and amplitude of the electrocardiogram are different, it can be roughly positioned for the standard electrocardiogram, but the error rate for the variable electrocardiogram is very high, and the robustness is poor.

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Electrocardio waveform positioning and classification model device adopting multi-scale visual field deep learning
  • Electrocardio waveform positioning and classification model device adopting multi-scale visual field deep learning
  • Electrocardio waveform positioning and classification model device adopting multi-scale visual field deep learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0036] This embodiment provides an ECG waveform positioning and classification model device using multi-scale field of view deep learning, such as figure 1 shown, including:

[0037] --Residual network module (ResNet network), used to receive the input of the ECG signal map and output the characteristic information of the ECG signal map; after the ResNet network, the structure of (BatchSize, 2048, 626, 1) can be obtained;

[0038] --Multi-scale aggregation modules, such as figure 2 As shown, it includes a pooling layer, several branches and a concat feature fusion layer. Each branch includes an adaptive pooling layer and a convolutional layer with different step lengths. Each branch can output feature variables of the same length and pass concat The feature fusion layer performs feature fusion; for example, the number of branches in the multi-scale positioning and classification module is 4. The step size of the pooling layer is 2, and each branch includes an adaptive pooli...

Embodiment 2

[0059] This embodiment is an improved ECG waveform positioning and classification model device using multi-scale field of view deep learning based on embodiment 1. The structure of the multi-scale aggregation module is different from that of embodiment 1, including:

[0060] The residual network module is used to receive the input of the electrocardiogram and output the characteristic information of the electrocardiogram;

[0061] Multi-scale aggregation module, including three pooling layers, 4 branches and 3 concat feature fusion layers, 4 branches including adaptive pooling layer and convolution layer with different step lengths, each branch can output the same length Feature variables, the steps of the three pooling layers are 2, 4 and 6 respectively, and the output of the pooling layer with a step size of 2 is connected to the adaptive pooling layer of the first branch, the second branch and the third branch , the output of the pooling layer with a stride of 4 is connecte...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

According to the method, a multi-scale deep network model is provided to learn multi-scale features for variable features of an electrocardiogram, the multi-scale features are fused, position featuresare well detected according to the fused features, and waveform positions and waveform categories can be outputted at a time.

Description

technical field [0001] The application belongs to the technical field of electrocardiographic signal processing, and in particular relates to an electrocardiographic waveform positioning and classification model device using multi-scale field of view deep learning. Background technique [0002] In the field of electrocardiography, some diagnostic procedures for premature atrial premature are to locate all heartbeat waves first, and then classify each heartbeat, based on the heartbeat classification and statistical results, the diagnostic conclusions such as occasional premature atrial premature, frequent premature atrial and bigeminy are given. . Due to the large number of waveforms and the limitation of human memory, this is a part that is most likely to be missed and misjudged in diagnosis for doctors. [0003] At present, the main application of deep learning in electrocardiographic diagnosis is the classification algorithm, which inputs a segment of electrocardiogram an...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
IPC IPC(8): A61B5/318A61B5/00
CPCA61B5/7267
Inventor 张德涛
Owner SHANGHAI SID MEDICAL CO LTD