Iterative Image Reconstruction Using HYPR Constraints
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Magnetic resonance imaging (MRI) reconstruction from partial k-space samples often results in artifacts due to under-sampling, with existing methods struggling to produce accurate and artifact-free images in a timely manner, especially when neighboring pixels have differing signal time courses.
Innovation Solution
The proposed method incorporates a constrained reconstruction process, using highly constrained projection reconstruction (HYPR) as a constraint in an iterative method like conjugate gradient (CG) to identify and correct inconsistencies between reference data and reconstructed images, thereby reducing artifacts and improving image quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If partial k-space acquisition is used to reduce scan time, then productivity is improved, but manufacturing precision deteriorates due to artifacts and image quality degradation
Solution Approach 1:
The method performs a preliminary highly-constrained reconstruction to generate an initial image estimate before the iterative refinement process. This preliminary action provides a starting point that incorporates anatomical constraints, enabling the iterative algorithm to converge faster and produce higher quality images from partial k-space data without requiring full sampling
Solution Approach 2:
The iterative reconstruction method uses feedback loops where the reconstructed image is continuously compared against the acquired k-space data, and correction factors are applied in successive iterations. This feedback mechanism allows the algorithm to progressively reduce artifacts and improve image quality while working with undersampled data
2Manufacturing precision
If iterative reconstruction methods are used to correct artifacts, then manufacturing precision is improved, but loss of time increases due to computational complexity
Solution Approach 1:
By performing a preliminary highly-constrained reconstruction before the iterative process, the method establishes a good initial estimate that is already close to the final solution. This preliminary action significantly reduces the number of iterative steps needed, thereby reducing total reconstruction time while maintaining high image accuracy
Solution Approach 2:
The method applies constraints selectively to specific regions or aspects of the image reconstruction rather than uniformly across all parameters. This partial application of constraints reduces computational burden while still achieving the necessary correction of artifacts and improvement of image accuracy
3Device complexity
If zero-filling is used to fill missing k-space data, then device complexity is reduced, but manufacturing precision deteriorates due to artifact generation
Solution Approach 1:
The highly-constrained reconstruction acts as an intermediary step between the raw partial k-space data and the final iterative refinement. This intermediary process generates an initial image estimate that incorporates anatomical constraints, providing a better foundation for the iterative algorithm and avoiding the artifact-prone zero-filling approach
Data Source
AI summary
Systems and methods using an image produced by a constrained image reconstruction process as a constraint in a forward iterative reconstruction process are described. One example system may include a constrained reconstruction logic to receive an initial data having an initial format and to produce an image data. The example system may include an iterative reconstruction logic that uses the image data as a constraint in a forward iterative step and that computes a correction factor based on comparing the image data to a reference data. The example system may include a deconstruction logic to deconstruct the image data into a deconstructed image data having the initial format and to selectively update the deconstructed image data based, at least in part, on the correction factor.