Data Consistency Operation for MRI Image Reconstruction

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing image reconstruction algorithms, particularly machine-learning based methods, face challenges in enforcing consistency between input images and reconstructed images, leading to artifacts and instability, especially in medical imaging applications.

Innovation Solution

The implementation of a data-consistency operation (DCO) that determines the contribution of source data from the input dataset to the K-space representation of the reconstructed image, ensuring that the final reconstructed image aligns with the input image, thereby reducing artifacts and improving stability.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If machine-learning based reconstruction algorithms are used, then image quality is improved, but consistency between input and reconstructed images deteriorates

Engineering Contradiction:

Improveimage qualityVSAvoidconsistency between input and reconstructed images

Core Design Contradiction:

Measurement precisionVSReliability

Solution Approach 1:

The patent implements a feedback mechanism by applying a data consistency operation (DCO) to the reconstructed image. The DCO compares the reconstructed image with the original input image and adjusts the reconstructed image to ensure consistency, thereby resolving the contradiction between improved image quality and maintained reliability

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The patent introduces an intermediary component (the data consistency operation) that mediates between the machine-learning reconstruction algorithm and the final output. This intermediary ensures that while the reconstruction algorithm improves image quality, the final reconstructed image remains consistent with the input data

Inventive Principle:

Principle #24Intermediary (Mediator)

2Measurement precision

If machine-learning based reconstruction algorithms are used, then image quality is improved, but artifacts and instability increase

Engineering Contradiction:

Improveimage qualityVSAvoidartifacts and instability

Core Design Contradiction:

Measurement precisionVSObject-affected harmful factors

Solution Approach 1:

The patent converts the potentially harmful effects of machine-learning artifacts into a beneficial process by using the data consistency operation to identify and correct these artifacts. The DCO leverages the reconstructed image while removing unwanted artifacts, thus converting the harmful side effect into a refinement step

Inventive Principle:

Principle #22Blessing in disguise (Convert harm into benefit)

3Reliability

If data consistency operation is applied, then consistency between input and reconstructed images is improved, but computational complexity increases

Engineering Contradiction:

Improveconsistency between input and reconstructed imagesVSAvoidcomputational complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The patent applies partial action by implementing the data consistency operation selectively rather than processing the entire image uniformly. The DCO focuses computational resources on specific regions or aspects of the image where consistency needs to be enforced, reducing overall computational complexity while maintaining reliability

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS12315047B2Data-consistency for image reconstruction

Publication Date: 2025.05.27 SIEMENS HEALTHINEERS AG

AI summary

A computer-implemented method includes, based on an input dataset defining an input image, determining a reconstructed image using a reconstruction algorithm, and executing a data-consistency operation for enforcing consistency between the input image and the reconstructed image. The data-consistency operation determines, for multiple K-space positions at which the input dataset comprises respective source data, a contribution of respective K-space values associated with the input dataset to a K-space representation of the reconstructed image.