HYPR Iterative Image Reconstruction for MRI Artifact Reduction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current image reconstruction methods in MRI and CT face challenges with streak artifacts due to insufficient sampling, particularly in clinical applications where scan time is critical, such as time-resolved angiography, and struggle to maintain high signal-to-noise ratio (SNR) and accuracy, especially when imaging moving subjects or non-sparse k-space data sets.

Innovation Solution

The HYPR method iteratively reconstructs images using an initial composite image, which is updated with each iteration, allowing for improved convergence to a high-quality image despite undersampling, and preserves phase information by performing calculations using complex numbers, effectively reducing artifacts and enhancing SNR.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If projection reconstruction methods are used with insufficient views to reduce scan time, then productivity is improved, but manufacturing precision deteriorates due to streak artifacts

Engineering Contradiction:
Improvescan speedVSAvoidimage quality
Core Design Contradiction:
ProductivityVSManufacturing precision

Solution Approach 1:

A composite image is reconstructed in advance from all available projection data before the actual image reconstruction. This preliminary composite image serves as a reference to guide the reconstruction of individual images from undersampled projections, preventing streak artifacts while maintaining fast scan speeds.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The composite image acts as an intermediary between the undersampled projection data and the final reconstructed image. By using the composite image to constrain and guide the reconstruction process, the method bridges the gap between limited data and high-quality images, eliminating streak artifacts without requiring additional views.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Manufacturing precision

If the number of views is increased to improve image quality, then manufacturing precision is improved, but productivity deteriorates due to longer scan time

Engineering Contradiction:
Improveimage qualityVSAvoidscan speed
Core Design Contradiction:
Manufacturing precisionVSProductivity

Solution Approach 1:

The composite image is constructed in advance from all projection data, capturing complete anatomical information. This preliminary step allows subsequent individual image reconstructions to use fewer views while maintaining high quality, as the composite image provides the missing information that would otherwise require additional views.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The composite image provides feedback information about the complete anatomy that is used to constrain and guide the reconstruction of individual images from undersampled projections. This feedback mechanism ensures high image quality without requiring a proportional increase in the number of views, thus maintaining fast scan speeds.

Inventive Principle:
Principle #23Feedback

3Productivity

If conventional reconstruction methods are used with undersampled data, then productivity is improved, but measurement precision deteriorates due to loss of phase information and low SNR

Engineering Contradiction:
Improvescan speedVSAvoidsignal-to-noise ratio
Core Design Contradiction:
ProductivityVSMeasurement precision

Solution Approach 1:

The composite image serves as an intermediary that preserves phase information from all projections. By using this composite image to guide the reconstruction process, the method recovers phase information that would be lost in conventional undersampled reconstructions, thereby maintaining high SNR and measurement precision while using fewer views.

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentEP1959397B1Iterative HYPR medical image reconstruction
Publication Date: 2019.08.07 WISCONSIN ALUMNI RES FOUND
  • EP1959397B1 patent drawingFigure 1
  • EP1959397B1 patent drawingFigure 2A~2B
  • EP1959397B1 patent drawingFigure 3

AI summary

An image reconstruction method includes reconstructing an initial composite image of a subject using a conventional reconstruction method. The initial composite image employs the best information available regarding the subject of the scan and this information is used to constrain the reconstruction of a highly undersampled or low SNR image frames. This highly constrained image reconstruction is repeated a plurality of iterations with the reconstructed image frame for one iteration being used as the composite image for the next iteration. The reconstructed image frame rapidly converges to a final image frame.