Eureka AIR delivers breakthrough ideas for toughest innovation challenges, trusted by R&D personnel around the world.

Multi-center child X-ray chest radiography image lung segmentation method based on TransUNet model

A multi-center, children's technology, applied in the field of medical artificial intelligence, can solve the problems that the accuracy and efficiency of segmentation need to be further improved, and the long-range relationship cannot be effectively modeled, so as to achieve good generalization, improvement of segmentation accuracy and efficiency, and good segmentation performance effect

Active Publication Date: 2022-07-05
ZHEJIANG UNIV
View PDF6 Cites 4 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] However, due to the inherent locality of convolution in these existing networks, convolution-based methods cannot effectively model long-range relationships, and the accuracy and efficiency of segmentation need to be further improved.

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
  • Multi-center child X-ray chest radiography image lung segmentation method based on TransUNet model
  • Multi-center child X-ray chest radiography image lung segmentation method based on TransUNet model
  • Multi-center child X-ray chest radiography image lung segmentation method based on TransUNet model

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0038] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be pointed out that the following embodiments are intended to facilitate the understanding of the present invention, but do not have any limiting effect on it.

[0039] like figure 1 As shown, a multi-center child X-ray chest X-ray image lung segmentation method based on the TransUNet model includes the following steps:

[0040] 1. Image preprocessing

[0041] Collect children's chest X-ray image data from multiple centers. Since the images come from multiple centers, the imaging standards are inconsistent, so grayscale truncation of the images is required. In addition, chest X-rays are skewed, and the chest is not completely centered on the image due to the size of the doctor's crop. Therefore, the steps of grayscale truncation are as follows: first obtain the image size (x, y), and then intercept the central area [x / 4:x*3 / 4,y / 4:y*3 / ...

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

The invention discloses a multi-center child X-ray chest radiography image lung segmentation method based on a TransUNet model. The method comprises the following steps: (1) collecting a multi-center child X-ray chest radiography image and preprocessing the multi-center child X-ray chest radiography image; (2) dividing the data into a training set, a verification set and a test set; (3) a segmentation model is constructed, a Transform layer is added to the segmentation model on the basis of UNet, and the segmentation model comprises four parts of three-time down-sampling, a linear layer, the Transform layer and up-sampling; (4) sending the training set into the constructed segmentation model for training, evaluating the performance of the segmentation model by using the verification set, adjusting the hyper-parameters of the model according to the evaluation effect, and finally obtaining the segmentation model with the performance reaching the standard through repeated training and verification; and (5) inputting a to-be-segmented multi-center child X-ray chest radiography image into the trained segmentation model so as to intelligently segment a lung region. The method provided by the invention combines the advantages of the Transformers network and the UNet network, and has relatively high segmentation precision and efficiency.

Description

technical field [0001] The invention belongs to the field of medical artificial intelligence, in particular to a lung segmentation method based on a TransUNet model of a multi-center children's X-ray chest X-ray image. Background technique [0002] Pneumonia in children is the top three diseases in children with infection rates. At present, many researchers use intelligent algorithms to diagnose pneumonia in children, and the pre-step of image-based diagnosis is the segmentation of lungs. Image-based lung classification also helps provide accuracy in lung nodule detection, bronchitis detection, and more. Conventional lung detection images for children mainly include CT and chest X-ray. The former has many slices and generally has higher segmentation accuracy, while the latter has only one image, and the quality of the images obtained on different machines is different, so lung segmentation is performed on them. more difficult. [0003] In response to this problem, many res...

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): G06T7/00G06T7/11G06N3/04G06N3/08
CPCG06T7/0012G06T7/11G06N3/08G06T2207/10116G06T2207/30061G06T2207/20081G06T2207/20084G06N3/045
Inventor 俞刚陈凌栋黄坚沈忱李竞孙晨升马晓辉余卓
Owner ZHEJIANG UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Eureka Blog
Learn More
PatSnap group products