MRI Motion Correction via K-Space Segmentation and Neural Network Estimation

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Magnetic Resonance Imaging (MRI) systems face challenges in correcting motion artifacts due to voluntary and involuntary subject movement, particularly when using Cartesian sampling patterns which have limited redundancy in k-space data, leading to image degradation.

Innovation Solution

The method involves dividing k-space data into multiple disjoint groups, selecting one as a reference, and using a spatial transformation estimation module, potentially implemented as a neural network, to calculate spatial transforms between these groups, enabling motion correction and reconstruction of corrected MRI images.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If Cartesian sampling pattern is used in k-space, then acquisition time is reduced and scanning efficiency is improved, but motion artifact correction capability deteriorates due to limited redundancy in k-space data

Engineering Contradiction:

Improvescanning efficiencyVSAvoidmotion artifact correction capability

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The k-space data is divided into multiple disjoint groups, each acquired at different time points. This segmentation allows independent motion estimation for each group relative to a reference group, enabling motion correction even with Cartesian sampling that lacks temporal overlap redundancy.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

An intermediate magnetic resonance image is reconstructed from each k-space data group to serve as a mediator for motion estimation. These intermediate images contain sufficient structural information to calculate spatial transforms between groups, bridging the gap between Cartesian sampling and motion correction requirements.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Reliability

If k-space data is divided into multiple disjoint groups, then motion correction is enabled, but data redundancy is reduced leading to potential loss of motion information

Engineering Contradiction:

Improvemotion correction capabilityVSAvoidmotion information

Core Design Contradiction:

ReliabilityVSLoss of information

Solution Approach 1:

Intermediate magnetic resonance images are reconstructed from each k-space data group to serve as intermediaries for motion estimation. These intermediate images preserve sufficient structural information to accurately calculate spatial transforms between groups, preventing loss of motion information despite data division.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

Traditional motion correction relying on k-space data redundancy is replaced with a neural network-based spatial transformation estimation system. The neural network processes intermediate images to calculate motion fields, substituting mechanical/data redundancy with intelligent information extraction.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Measurement precision

If neural network is used for spatial transformation estimation, then motion correction accuracy is improved, but computational complexity and processing time increase

Engineering Contradiction:

Improvemotion correction accuracyVSAvoidcomputational complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

Intermediate magnetic resonance images are reconstructed from k-space data groups before spatial transformation estimation. This preliminary action prepares the data in a form that contains sufficient structural information for accurate motion estimation, reducing the computational burden on the neural network while maintaining precision.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system uses itself to correct motion artifacts by comparing each k-space data group against a reference group through intermediate image reconstruction and spatial transformation estimation. This self-referential approach enables accurate motion correction without requiring external reference data or complex additional hardware.

Inventive Principle:

Principle #25Self-service

Data Source

PatentUS11519991B2Motion estimation and correction in magnetic resonance imaging

Publication Date: 2022.12.06 KONINKLIJKE PHILIPS NV

AI summary

A method of medical imaging including receiving k-space data that is divided into multiple k-space data groups, selecting one of the multiple k-space data groups as a reference k-space data group, and calculating spatial transform data for each of the multiple k-space data groups by inputting the multiple k-space data groups and the reference k-space data group into a transformation estimation module. The spatial transformation estimation module is configured for outputting spatial transform data descriptive of a spatial transform between a reference k-space data group and multiple k-space data groups in response to receiving the reference k-space data group and the multiple k-space data groups as input. The method further comprises reconstructing a corrected magnetic resonance image according to the magnetic resonance imaging protocol using the multiple k-space data groups and the spatial transform data for each of the multiple k-space data groups.