MRI K-Space Segment Merging for Artifact-Resistant Reconstruction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing magnetic resonance imaging (MRI) techniques using readout-segmented echo-planar imaging (RS-EPI) face challenges in accurately reconstructing images due to artifacts from susceptibility and B0 inhomogeneity, motion, and phase variations, which affect the quality and diagnostic accuracy of clinical MR images.

Innovation Solution

An iterative MRI reconstruction method that includes pre-processing each segment of K-space data using echo planar imaging (EPI) Nyquist ghost correction and parallel acquisition, followed by phase correction and synthesis, and iteratively merging segments using entropy minimization and k-space center correction to generate a complete K-space data set for improved image reconstruction.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If parallel imaging techniques are used to reduce scan time, then productivity is improved, but measurement precision deteriorates due to undersampling of k-space data

Engineering Contradiction:
Improvescan timeVSAvoidimage quality
Core Design Contradiction:
ProductivityVSMeasurement precision

Solution Approach 1:

The system uses auto-calibration signals to automatically determine the sensitivity profiles of receiver coils, providing feedback that enables accurate image reconstruction from undersampled data without manual calibration

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The method performs self-calibration by incorporating calibration signals within the imaging sequence itself, eliminating the need for separate calibration procedures and enabling the system to automatically compensate for undersampling effects

Inventive Principle:
Principle #25Self-service

2Loss of time

If k-space is undersampled to accelerate imaging, then scan time is reduced, but image quality deteriorates due to aliasing artifacts

Engineering Contradiction:
Improvescan timeVSAvoidimage quality
Core Design Contradiction:
Loss of timeVSManufacturing precision

Solution Approach 1:

Sensitivity profiles derived from auto-calibration signals serve as intermediary information that enables separation of aliased signals during image reconstruction, allowing high-quality images from undersampled data

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The system performs preliminary auto-calibration to establish sensitivity profiles before reconstructing the final image, preparing necessary information in advance to handle undersampling effects

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP3749185B1A method and an apparatus for reconstructing a magnetic resonance image
Publication Date: 2026.05.20 SAMSUNG ELECTRONICS CO LTD
  • EP3749185B1 patent drawingFigure 1a~1c
  • EP3749185B1 patent drawingFigure 1d~3
  • EP3749185B1 patent drawingFigure 4~5

AI summary

Provided is an apparatus of reconstructing a magnetic resonance (MR) image, the apparatus including: a memory storing instructions; and at least one processor configured to execute the instructions to: obtain a plurality of segments of K-space data corresponding to a plurality of pulses which are applied to an object based on a pulse sequence; determine, based on radio frequency (RF) coils of the apparatus, a correction coefficient for merging the plurality of segments of K-space data; and generate a magnetic resonance (MR) image of the object by merging the plurality of segments of K-space data based on the determined correction coefficient.