Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Brain tumor segmentation algorithm based on UNet++ optimization and weight budget

A segmentation algorithm, brain tumor technology, applied in the field of brain tumor segmentation algorithm, can solve the problems of low segmentation accuracy, incomplete semantic information of UNet network, slow training speed, etc., achieve good segmentation effect and solve the effect of slow network training speed

Inactive Publication Date: 2021-04-02
CHANGCHUN UNIV OF TECH
View PDF6 Cites 7 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] The technical problem solved by the present invention is aiming at the problem of incomplete semantic information and loose connection of contextual information in UNet network and the problem of low segmentation accuracy of internal structure of brain tumor in 3D brain tumor segmentation, and proposes a method based on UNet++ optimization and weight budget The brain tumor segmentation algorithm uses the UNet++ network model to make the semantic information of the UNet network incomplete and the context information not closely connected. To a certain extent, its tight network structure is solved, so that the internal tissue segmentation effect of the 3D brain tumor segmentation is better. good
The weight budget solves the problem of slow training speed of the UNet++ network model due to the complex network structure, and at the same time, the segmentation accuracy will be better

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
  • Brain tumor segmentation algorithm based on UNet++ optimization and weight budget
  • Brain tumor segmentation algorithm based on UNet++ optimization and weight budget
  • Brain tumor segmentation algorithm based on UNet++ optimization and weight budget

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0060] For those skilled in the art, it is understandable that some well-known structures and descriptions thereof may be omitted in the drawings. The technical solutions of the present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0061] The present invention provides a brain tumor segmentation algorithm based on UNet++ optimization and weight budgeting. The method realizes the segmentation of the whole tumor, tumor core and enhanced tumor core of brain tumors, and provides high-precision segmentation of brain tumor MRI images. Repeatable measurements and evaluations provide more accurate tumor image segmentation maps.

[0062] figure 1 It is the flow chart of the method of the present invention, first image preprocessing, BraTS2018 and BraTS2019 become the input required by the network; then build an improved UNet network model, and use it to train the data, and save the network weight with the best effect; Next,...

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

For brain tumor magnetic resonance imaging (MRI) multi-mode accurate segmentation, compared with a 2D brain tumor segmentation network, a 3D brain tumor segmentation network can better store interlayer information and has a better segmentation effect in the learning process, but is always poor in effect when tumor internal tissues are segmented. An UNet++ networkimproved based on a UNet network isintroduced, and the characteristic of tight connection of the network structure is utilized, so thatthe segmentation precision of brain tumor internal tissues is improved; a residual module is addedto solve the problems of information loss and degradation during network training, so that a proper network structure is constructed; and however, the constructed network is complex in structure and slow in training time, and on the basis, the weight budget is utilized, so that the segmentation precision is improved while the problem of slow training is solved. Experimental results show that the improved network has a good segmentation effect on the interior of the brain tumor and has better performance than a typical brain tumor segmentation method.

Description

technical field [0001] The present invention proposes a brain tumor segmentation algorithm based on deep learning, which adopts a brain tumor segmentation algorithm based on UNet++ optimization and weight budget. The improved UNet++ network model can be used to more accurately segment MRI images of brain tumors. It can more accurately segment the internal tissues of brain tumors while ensuring the overall segmentation accuracy. The use of weight budget can solve the problem of slow training time caused by the complex structure of UNet++ network. At the same time, it is better than direct training for internal tissue segmentation of brain tumors. Background technique [0002] At present, hospitals generally use manual delineation to determine the radiotherapy target area when formulating radiotherapy plans. However, manual delineation has many disadvantages: on the one hand, the doctor’s screening process is very time-consuming; on the other hand, due to the invasive growth ...

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): G06T7/00G06T7/10G06N3/04
CPCG06T7/0012G06T7/10G06T2207/10088G06T2207/20081G06T2207/30016G06N3/045
Inventor 侯阿临吴浪孙弘建杨骐豪崔博姬鹏季鸿坤刘丽伟李秀华梁超杨冬
Owner CHANGCHUN UNIV OF TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products