Parallel MR Reconstruction Using Coil Sensitivity Maps for Aliasing Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing MR image reconstruction techniques face challenges in effectively correcting aliasing artifacts in position space for parallel MR imaging with regular undersampling, particularly when using machine learning models, as they do not fully leverage the information from reconstruction weights and coil sensitivity maps.
Innovation Solution
A computer-implemented method for MR image reconstruction that utilizes effective coil sensitivity maps, determined from reconstruction weights, to correct aliasing in position space, combining techniques like SENSE and machine learning models to enhance image quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If machine learning models are used for MR image reconstruction with regular undersampling, then reconstruction speed and productivity are improved, but aliasing artifact correction performance deteriorates compared to k-space based methods
Solution Approach 1:
The patent introduces effective coil sensitivity maps as an intermediary that bridges machine learning reconstruction and aliasing correction. These maps are determined from reconstruction weights and serve as a mediator to correct aliasing artifacts in position space, enabling ML models to achieve GRAPPA-level artifact correction while maintaining fast reconstruction speeds.
Solution Approach 2:
The patent transitions the aliasing correction from k-space domain to position space domain. By determining effective coil sensitivity maps in position space and using them to correct aliasing artifacts after initial ML reconstruction, the method operates in a different dimensional space, achieving comparable artifact correction to k-space methods like GRAPPA while benefiting from ML reconstruction speed.
2Productivity
If higher acceleration factors are used in parallel MR imaging, then scanning time is reduced and productivity increases, but image quality deteriorates due to increased aliasing artifacts
Solution Approach 1:
Effective coil sensitivity maps determined from reconstruction weights serve as an intermediary to correct aliasing artifacts. This allows higher acceleration factors to be used during data acquisition while the intermediate correction step restores image quality, decoupling the trade-off between scanning speed and image quality.
Solution Approach 2:
The method performs preliminary determination of effective coil sensitivity maps from the undersampled data itself, before final image reconstruction. This preliminary action enables the system to prepare correction information in advance, allowing higher acceleration factors without quality loss.
3Ease of operation
If conventional reconstruction techniques are used without leveraging reconstruction weights effectively, then implementation simplicity is maintained, but aliasing correction performance in position space deteriorates
Solution Approach 1:
The patent introduces effective coil sensitivity maps as an intermediary component that can be determined from existing reconstruction weights using established methods. This intermediary enables position space aliasing correction without requiring completely new reconstruction algorithms, maintaining implementation simplicity while improving performance.
Solution Approach 2:
The method changes the parameter representation by determining effective coil sensitivity maps from reconstruction weights rather than using traditional approaches. This parameter transformation enables better aliasing correction in position space while building upon conventional reconstruction frameworks.
Data Source
AI summary
Techniques are provided for image reconstruction in parallel MR imaging, in which a respective set of regularly undersampled MR measurement data in k-space representing an imaged object is received for each of a plurality of coil channels. For each pair of coil channels of the plurality of coil channels, a respective set of reconstruction weights for reconstructing MR data at k-space points, which are not measured according to the undersampling, from the MR measurement data, is received. For each of the plurality of coil channels, a respective coil sensitivity map is determined depending on the respective sets of reconstruction weights for the respective coil channel. A reconstructed MR image is generated based on the coil sensitivity maps.