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

Optimized regularization and CNN-based undersampled magnetic resonance image high-performance reconstruction method

A magnetic resonance image and undersampling technology, which is applied in the field of image processing, can solve the problems of long magnetic resonance imaging time and good reconstruction performance, and achieve the effects of fast imaging time, auxiliary judgment, and reduced scanning time

Pending Publication Date: 2022-07-15
NANJING MEDICAL UNIV
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to solve the problem of too long magnetic resonance imaging time, and provide a high-performance reconstruction method of under-sampled magnetic resonance images based on regularization and CNN. The method model has the advantages of portability, end-to-end, etc., and the reconstruction performance is good

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
  • Optimized regularization and CNN-based undersampled magnetic resonance image high-performance reconstruction method
  • Optimized regularization and CNN-based undersampled magnetic resonance image high-performance reconstruction method
  • Optimized regularization and CNN-based undersampled magnetic resonance image high-performance reconstruction method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0051] The present invention will be described in detail below with reference to specific embodiments.

[0052] The present invention proposes a high-performance reconstruction method of under-sampling magnetic resonance images based on optimal regularization and CNN, which includes the following steps.

[0053] Step 1: Introduce a general model to solve the most sparse solution of undersampled MR images. The general formula is:

[0054]

[0055] where I∈C N is the reconstructed MR image with N pixels. y∈C M is the acquired undersampled k-space MR image data. K is the observation matrix, expressed as:

[0056] K=MF (2)

[0057] where M is the diagonal matrix of the undersampled mask in K-space and F is the 2D discrete Fourier transform.

[0058] Since the N dimension of the sparse coefficient θ is much larger than the M dimension of the observation matrix K, the reconstruction of the original signal can be regarded as a L0 norm minimization problem of NP, but the non-...

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 regularization and CNN-based high-performance reconstruction method for an undersampled magnetic resonance image, and the method employs a high-quality magnetic resonance image as a training set, and designs and trains an offline convolutional neural network to map a low-quality MR image obtained from undersampled k-space data and the high-quality image. The network model is portable, end-to-end operation is carried out, the imaging time is short, the reconstruction effect can display more local details, the trained model can process a low-sampling-rate magnetic resonance image newly scanned by a patient, the problem that the magnetic resonance scanning time is too long is solved, and doctors are assisted in judging the illness state.

Description

technical field [0001] The invention relates to a high-performance reconstruction method of under-sampling magnetic resonance images based on optimal regularization and CNN, and belongs to the field of image processing. Background technique [0002] The magnetic resonance imaging time is too long, which will bring many clinical problems. The current research direction is to reduce the sampling rate of K-space to speed up the scanning time. artifacts. Since the advent of compressive sensing in 2004, compressive sensing technology reconstructs medical images without distortion with fewer k-space samples, because the samples that need to be captured are greatly reduced, thereby speeding up the imaging time. [0003] Because the similarity between the same organ and tissue information in different human bodies is not considered, there are many redundant calculations in the algorithm. Deep learning has developed rapidly in the field of medical images in recent years. Through a ...

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): G06T11/00A61B5/00A61B5/055G06F17/14G06F17/16G06N3/04G06N3/08
CPCG06T11/008A61B5/055A61B5/0033A61B5/7257A61B5/7264A61B5/7267G06F17/141G06F17/16G06N3/08G06N3/045
Inventor 王伟冯锐吴小玲李修寒曹达
Owner NANJING MEDICAL 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