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

A method for Parkinson's disease classification and lesion area labeling in MRI images

A Parkinson's disease and image technology, applied in the field of medical images of computer vision, can solve the problems of inaccurate lesion area and poor classification results, and achieve the effects of improving training efficiency, reducing dimensions, and reducing gradient disappearance.

Active Publication Date: 2022-04-26
SOUTHWEST JIAOTONG UNIV
View PDF7 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Therefore, using the existing model to classify and label MRI images, the classification results obtained are poor, and the labeled lesion area is inaccurate

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
  • A method for Parkinson's disease classification and lesion area labeling in MRI images
  • A method for Parkinson's disease classification and lesion area labeling in MRI images
  • A method for Parkinson's disease classification and lesion area labeling in MRI images

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0044] Specific examples of the present invention are given below to further illustrate the present invention.

[0045] S1: Download the medical image dataset, extract 1557 PD (Parkinson's disease) images, 543 Control (normal) images, and 193 Prodromal (latency) images according to the label documents.

[0046] S2: Preprocess the MRI images. For images that do not meet the resolution, use the cubic interpolation method and average pooling to perform upsampling and downsampling respectively, and then obtain the training set and save it in tfrecord format.

[0047] Downsampling uses the average pooling method, that is, only averages the feature points in the neighborhood. The formula is:

[0048] alpha i ∈{0,1},

[0049]

[0050]

[0051] Upsampling uses cubic interpolation.

[0052] S3: Construct classification module, such as figure 1 shown. The input is a tensor of [224,224,3], through the first convolutional layer (conv), the convolution kernel size is 7*7, and th...

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 method for classifying Parkinson's disease of MRI images and labeling lesion areas, comprising the steps of: constructing a Parkinson's disease classification model of MRI images; constructing a Parkinson's disease lesion labeling model; using the Parkinson's disease classification model of MRI images , to classify the MRI image to be tested; if the classification is determined to be a Parkinson's disease MRI image or a Parkinson's disease latency MRI image, the lesion area is marked using the Parkinson's disease lesion labeling model. Wherein, constructing the Parkinson's disease classification model of the MRI image includes the steps of constructing the classification module and inputting the training set into the classification module, and updating network parameters through backpropagation to obtain the MRI image classification model. The beneficial effects of the present invention are: using the method of classifying first and then labeling, on the basis of improving the accuracy of the classification model, the location of the lesion area is more accurate and effective, and unnecessary training is avoided at the same time, and the training efficiency is greatly improved. The model is more robust.

Description

technical field [0001] The invention relates to the field of medical images of computer vision, in particular to a method for classifying Parkinson's disease of MRI images and marking lesion areas. Background technique [0002] With the development of medical imaging technology, CT, MRI and other medical imaging technologies have become an important basis and means for the diagnosis of Parkinson's disease. However, in the existing technology, it is necessary to rely on experts to analyze and confirm medical images, which not only consumes a lot of labor and is inefficient, but also because the imaging machines of each hospital are different, the gap in hardware will make the image of the same part of the same person There will also be great differences, which will have a great impact on manual analysis. [0003] With the rapid development of computer vision, more and more people combine computer vision with medical diagnosis, but there are still many problems in applying MR...

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 Patents(China)
IPC IPC(8): G06T7/00G06T7/70G06K9/62G06N3/04G06N3/08G06V10/764G06V10/82
CPCG06T7/0012G06T7/70G06N3/084G06T2207/10088G06V2201/03G06N3/045G06F18/24
Inventor 张晓博张哲浩李伟
Owner SOUTHWEST JIAOTONG UNIV
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

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com