Compressed Sensing ROI Segmentation for MRI Reconstruction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Magnetic Resonance Imaging (MRI) reconstruction using compressed sensing techniques faces challenges in reconstructing images under time-sensitive conditions with less than ideal data sampling, leading to inefficiencies due to the need for oversampling and time-consuming non-linear iterative optimization, especially when processing background noise.
Innovation Solution
The method involves separating the image into a region of interest (ROI) and a background region, applying compressed sensing only to iteratively update the ROI while maintaining the background, using a cost function that includes a fidelity term and a sparsity term with masks to retain or filter out the background, thereby reducing the data to be processed and reconstruction time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If compressed sensing is applied to the entire image including background regions, then image reconstruction can be performed, but reconstruction time increases significantly due to processing unnecessary background data
Solution Approach 1:
The patent divides the image into a region of interest (ROI) and a background region, applying compressed sensing only to the ROI. This segmentation allows the system to process only the necessary data (ROI) rather than the entire image, significantly reducing reconstruction time and computational burden while maintaining image quality in the critical area.
Solution Approach 2:
The patent extracts and removes the background region from the compressed sensing processing. By taking out the background data that would otherwise be processed unnecessarily, the system reduces the amount of data to be processed and accelerates reconstruction without compromising the quality of the ROI.
2Reliability
If oversampling is performed in radial trajectory acquisition, then aliasing artifacts are avoided, but reconstruction time increases due to non-linear iterative optimization of extra data
Solution Approach 1:
The patent extracts and removes the oversampled background regions from the reconstruction process. By taking out the peripheral background data that causes time-consuming optimization, the system maintains reliable aliasing artifact avoidance in the ROI while significantly reducing reconstruction time.
Solution Approach 2:
The patent segments the image data into ROI and background, applying reconstruction algorithms only to the ROI. This segmentation allows the system to maintain the reliability benefits of oversampling in the critical area while avoiding the time cost of processing the entire oversampled dataset.
3Manufacturing precision
If the entire image is processed using compressed sensing, then complete image reconstruction is achieved, but computational resources are wasted on background noise
Solution Approach 1:
The patent extracts and removes the background region from computational processing. By taking out the background noise that would consume computational resources, the system achieves efficient energy utilization while maintaining high reconstruction quality in the ROI through targeted processing.
Solution Approach 2:
The patent applies different processing quality to different regions: high-quality compressed sensing reconstruction is applied locally to the ROI where image quality matters, while the background region is handled differently (maintained or simplified). This local quality approach optimizes computational energy by concentrating resources where they provide the most value.
Data Source
AI summary
A method for reconstructing an image includes acquiring raw image data during a scan of an area, estimating an image from the raw image data, separating the estimated image into a region of interest (ROI) and a background region, and applying compressed sensing to iteratively update only the ROI and maintain the background region to reconstruct an image.