MRI Image Reconstruction from Sub-Sampled Signals
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Magnetic resonance imaging (MRI) techniques suffer from long imaging times, which can be burdensome for patients and hinder diagnosis, particularly in cases of claustrophobia, and there is a need for improved image quality and reduced acquisition time.
Innovation Solution
A magnetic resonance image processing apparatus and method utilizing a sub-sampled magnetic resonance signal, combined with parallel imaging techniques and artificial neural networks to accelerate image acquisition and enhance image quality, involving preprocessing and reconstruction using sensitivity matrices and neural network models to generate high-quality images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If MRI imaging time is reduced by sub-sampling, then productivity is improved, but measurement precision deteriorates
Solution Approach 1:
An artificial neural network is introduced as an intermediary between sub-sampled k-space data and the final image. The neural network learns to reconstruct high-quality images from incomplete data by mapping sub-sampled k-space to full-resolution images, effectively bridging the gap between fast acquisition and high quality output
Solution Approach 2:
The system performs preliminary parallel imaging reconstruction on sub-sampled data to generate initial images before applying the neural network. This preliminary action creates a foundation that the neural network then refines, combining the speed of parallel imaging with the quality enhancement of deep learning
2Productivity
If parallel imaging technique is used to accelerate acquisition, then productivity is improved, but manufacturing precision deteriorates
Solution Approach 1:
The neural network serves as a mediator that corrects the inherent inaccuracies of parallel imaging reconstruction. It takes the rapidly acquired but imperfect parallel imaging results and transforms them into high-fidelity images, maintaining both speed and accuracy
Data Source
AI summary
The present invention is to provide a magnetic resonance image processing method. According to an embodiment of the present invention, a magnetic resonance image processing method by a magnetic resonance image processing apparatus comprises the steps of: obtaining a sub-sampled magnetic resonance signal; acquiring first k-space data from the sub-sampled magnetic resonance signal using a first parallel imaging technique; obtaining a first magnetic resonance image from the first k-space data by using an inverse Fourier operation; generating first input image data by preprocessing the first magnetic resonance image; and obtaining a first output magnetic resonance image from the first input image data using a first artificial neural network mode.