Mammary gland nuclear magnetic image tumor segmentation method and device based on weak supervised learning

A weakly supervised, imaging technology, applied in the field of medical image processing, can solve the problem of pixel-level time-consuming, and achieve the effect of ensuring accuracy and improving efficiency

Pending Publication Date: 2020-08-21
BEIJING INSTITUTE OF TECHNOLOGYGY
View PDF4 Cites 10 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

All of these require complete annotation information to train the segmentation model, but manual annotation at the pixel level is time-consuming

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
  • Mammary gland nuclear magnetic image tumor segmentation method and device based on weak supervised learning
  • Mammary gland nuclear magnetic image tumor segmentation method and device based on weak supervised learning
  • Mammary gland nuclear magnetic image tumor segmentation method and device based on weak supervised learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0022] Such as figure 1 As shown, this method of breast MRI tumor segmentation based on weakly supervised learning includes the following steps:

[0023] (1) Mark the largest tumor section in the MRI image as a segmentation label;

[0024] (2) Perform grayscale normalization and resolution reduction processing on MRI images;

[0025] (3) Input the processed MRI image and the segmentation label into the deep learning neural network, combine the volume prediction weak supervision loss, train the deep learning neural network, and obtain the network model;

[0026] (4) input the image to be segmented into the network model for segmentation, and obtain the segmentation result;

[0027] (5) Based on the segmentation result of step (4), select the largest connected domain to remove noise, and obtain the final segmentation result.

[0028] In the present invention, by marking the largest tumor section in the MRI image as a segmentation label, the processed MRI image and the segment...

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 mammary gland nuclear magnetic image tumor segmentation method and device based on weak supervised learning, the incomplete annotation information can be utilized to train a segmentation model, and the segmentation efficiency is greatly improved while the precision is ensured. The method comprises the following steps: (1) marking a maximum tumor section in an MRI image as a segmentation label; (2) performing gray normalization and resolution reduction processing on the MRI image; (3) inputting the processed MRI image and the segmentation label into a deep learning neural network, and training the deep learning neural network in combination with volume prediction weak supervision loss to obtain a network model; (4) inputting a to-be-segmented image into the network model for segmentation to obtain a segmentation result; and (5) based on the segmentation result in the step (4), selecting a maximum connected domain to remove noise to obtain a final segmentation result.

Description

technical field [0001] The present invention relates to the technical field of medical image processing, in particular to a method for tumor segmentation of breast nuclear magnetic imaging based on weakly supervised learning, and a device for segmenting breast nuclear magnetic image tumors based on weakly supervised learning. Background technique [0002] Breast cancer is one of the most common malignant tumors affecting women's health. Early detection and early treatment can effectively reduce the mortality rate of breast cancer. At present, X-ray, ultrasound and magnetic resonance are the three most common ways to detect tumors. Among them, Magnetic Resonance Imaging (MRI) has rich three-dimensional information and high soft tissue resolution, so it is more and more used in the clinical diagnosis of breast cancer. [0003] One of the key steps in the diagnosis and treatment of breast tumors is the accurate segmentation of breast tumors. Manual annotation is time-consumi...

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
Patent Type & Authority Applications(China)
IPC IPC(8): G06T7/10G06T5/00G06N3/08
CPCG06T7/10G06T5/002G06N3/08G06T2207/10088G06T2207/20081G06T2207/30096
Inventor 杨健范敬凡王涌天孟宪琦
Owner BEIJING INSTITUTE OF TECHNOLOGYGY
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
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap