Dynamic MRI Reconstruction via Motion-Guided Block Low-Rank Sparsity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current MRI acceleration techniques, such as compressed sensing, face challenges in handling patient motion during dynamic imaging, leading to artifacts in images due to respiratory or other movements, which affects the spatiotemporal redundancy of data and reduces image quality.
Innovation Solution
The method involves acquiring undersampled MRI data, separating images into regions, performing motion tracking, grouping these regions into clusters based on spatial content, and applying a sparsity transform using singular value decomposition to exploit regional spatiotemporal sparsity, thereby compensating for motion and improving image reconstruction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If compressed sensing is used to accelerate dynamic MRI, then spatial resolution and temporal resolution are improved, but patient motion causes image artifacts and reduces image quality
Solution Approach 1:
The image is divided into multiple image regions that are processed separately. Motion tracking is performed for each region independently, and sparsity transforms are applied to clustered regions. This segmentation allows the system to handle motion in different parts of the image differently, reducing artifacts while maintaining high resolution.
Solution Approach 2:
Motion tracking is performed throughout the set of estimated images to generate motion data, which is then fed back into the reconstruction process. The motion compensation step uses this feedback to correct for patient motion, thereby reducing artifacts and improving image quality while maintaining the acceleration benefits of compressed sensing.
2Loss of time
If compressed sensing is used to accelerate dynamic MRI, then scanning time is reduced, but patient motion during scanning leads to artifacts
Solution Approach 1:
Motion tracking is performed throughout the acquisition of undersampled MRI data, before the final image reconstruction is completed. This preliminary motion assessment allows the system to prepare motion compensation strategies in advance, correcting for patient motion that occurs during the accelerated scan, thereby maintaining image quality while reducing scanning time.
3Reliability
If motion compensation is applied to correct for patient motion, then image quality is improved, but the complexity of the reconstruction algorithm increases
Solution Approach 1:
Instead of applying uniform motion compensation across the entire image, the system performs motion tracking and applies sparsity transforms to specific clustered image regions. This local approach targets motion-affected areas specifically, improving image quality while reducing the overall computational complexity compared to global motion compensation methods.
Solution Approach 2:
The image is divided into multiple regions that are processed separately through motion tracking and clustering. By segmenting the reconstruction process, the system manages complexity through modular processing of smaller region clusters rather than handling the entire image as a single complex unit.
4Productivity
If undersampled MRI data is acquired to reduce scanning time, then productivity is improved, but data redundancy is reduced making reconstruction more difficult
Solution Approach 1:
The system changes the sampling parameters by acquiring undersampled MRI data at accelerated rates. Combined with region-based clustering and sparsity transforms, this parameter change allows efficient use of reduced data while maintaining reconstruction quality through intelligent processing of the available information.
Solution Approach 2:
Multiple clustered image regions are merged through sparsity transforms to form a complete reconstructed image. By combining information from multiple clustered regions that have been processed individually, the system recovers lost redundancy and achieves high-quality reconstruction from undersampled data.
Data Source
AI summary
Some aspects of the present disclosure relate to systems and methods for accelerated dynamic magnetic resonance imaging (MRI). In an example embodiment, a method includes acquiring undersampled MRI data corresponding to a set of images associated with an area of interest of a subject, and separating an image of the set of images into image regions. The method also includes performing motion tracking for each of the image regions, grouping the motion-tracked image regions into clusters, and applying a sparsity transform to the clusters, to form sparsity-exploited, transformed image regions. The method further includes forming a set of merged images from the plurality of sparsity-exploited, transformed image regions, and updating the set of merged images based on data fidelity, to form an updated set of estimated images.