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

Method using convolution nerve network to remove magnetic resonance image downsampling artifact

A technology of convolutional neural network and magnetic resonance image, applied in biological neural network model, neural architecture, image enhancement, etc., can solve the problems of residual artifacts, low image quality, loss of detailed structure, etc., and achieve high resolution and Contrast, high robustness effects

Active Publication Date: 2017-12-22
SOUTHERN MEDICAL UNIVERSITY
View PDF4 Cites 26 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

This method is applicable to Cartesian sampling and Radial / Spiral sampling at the same time, but there will still be loss of detail structure and residual artifacts, and the image quality is not high

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
  • Method using convolution nerve network to remove magnetic resonance image downsampling artifact
  • Method using convolution nerve network to remove magnetic resonance image downsampling artifact
  • Method using convolution nerve network to remove magnetic resonance image downsampling artifact

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0062] Such as figure 1 As shown, based on the method of convolutional neural network to remove magnetic resonance image downsampling artifacts, the magnetic resonance image with artifacts is processed by the convolutional neural network to obtain an artifact-free result image.

[0063] The sample magnetic resonance image with artifacts is taken as the input image, processed by the convolutional neural network framework to obtain the optimal model, and then the magnetic resonance image to be processed is input into the optimal model to obtain the result image without artifacts.

[0064] Specific steps are as follows:

[0065] (1) Obtain the optimal model

[0066] T1. Preprocessing the magnetic resonance image of the sample.

[0067] T2. Build the basic framework of convolutional neural networks.

[0068] T3. Optimize the basic frame parameters of the convolutional neural network through the training data to obtain the optimal model.

[0069] (2) De-artifacting the image to be processed th...

Embodiment 2

[0103] Based on the method of convolutional neural network to remove magnetic resonance image downsampling artifacts, the other features are the same as in embodiment 1, except that: when i=4, the basic structure of the convolutional neural network is constructed in step T2 through the construction feature extraction in turn Layer, build feature enhancement layer, build nonlinear mapping layer and build reconstruction layer.

[0104] Build the feature extraction layer: extract features from the input data to get the first layer of output data.

[0105] Build a feature enhancement layer: further extract features from the output data of the first layer to obtain the output data of the second layer.

[0106] Build a non-linear mapping layer: map the output data group of the second layer to a sample reference image without artifacts to obtain the output data of the third layer.

[0107] Build the reconstruction layer: reorganize the output data of the third layer to output the predicted i...

Embodiment 3

[0115] The method based on the convolutional neural network to remove the magnetic resonance image down-sampling artifacts, other features are the same as the first embodiment, the difference lies in: Figure 3-4 As shown, image 3 The actual image of the artifact removed by the method of the present invention is given, including the reference image of the sample, the magnetic resonance image of the sample and the result image after the method of the present invention is used to remove the artifact, and the residual image relative to the reference image. by image 3 It can be seen that the method of the present invention can effectively and robustly remove Gibbs artifacts, retain detailed information, and improve image quality.

[0116] Figure 4 for image 3 It can be seen from the partially enlarged view that the method of the present invention can effectively remove the Gibbs artifact.

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

A method using the convolution nerve network to remove magnetic resonance image downsampling artifacts is provided; the method uses a sample magnetic resonance image with the artifact as an input image, and uses the convolution nerve network framework to process the image, thus finally obtaining a result image with no artifact; the method comprises the following steps: 1, obtaining an optimal model, T1, preprocessing the sample magnetic resonance image; T2, building the convolution nerve network basic framework; T3, initializing the convolution nerve network basic framework parameters; T4, using training data to optimize the convolution nerve network basic framework parameters, thus obtaining an optimal model; 2, using the optimal model to remove artifacts of the to-be-processed image so as to obtain a target image, T5, substituting the preprocessed to-be-processed image into the optimal model, and outputting the result image. The optimal model can effectively remove the artifacts caused by downsampling, thus obtaining high resolution and contract ratio, and well keeping image details.

Description

Technical field [0001] The present invention relates to the technical field of medical equipment, in particular to a method for removing down-sampling artifacts of magnetic resonance images based on convolutional neural networks. Background technique [0002] Magnetic resonance imaging is an important part of the field of medical diagnosis. However, artifacts caused by downsampling often appear in magnetic resonance imaging, such as Gibbs artifacts caused by the loss of high-frequency data due to limited sampling of k-space, and aliasing artifacts caused by downsampling in parallel imaging. As well as noise, fine streak artifacts that appear in the image during Radial / Spiral scanning, etc. The artifacts in the image caused by the drop-down will cause the image contrast and spatial resolution to be greatly reduced, so it is very necessary to find an effective method to remove the artifacts. [0003] In the past many years, many image post-processing algorithms to remove downsampli...

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): G06T5/00G06N3/04
CPCG06T2207/10088G06N3/045G06T5/80
Inventor 冯衍秋张倩倩
Owner SOUTHERN MEDICAL UNIVERSITY
Features
  • Generate Ideas
  • Intellectual Property
  • Life Sciences
  • Materials
  • Tech Scout
Why Patsnap Eureka
  • Unparalleled Data Quality
  • Higher Quality Content
  • 60% Fewer Hallucinations
Social media
Patsnap Eureka Blog
Learn More

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

© 2025 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com