Automated MRI Regularization Parameter Determination

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Magnetic resonance imaging (MRI) techniques face challenges in reducing measurement times due to limitations imposed by the Nyquist theorem, leading to increased acquisition times and potential aliasing artifacts, which are not effectively addressed by existing methods like parallel imaging and compressed sensing that require manual parameter tuning.

Innovation Solution

A method for generating MRI data sets using a turbo spin echo (TSE) sequence with automated determination of the regularization parameter λ, utilizing echo signals from a phase correction measurement to simplify compressed sensing reconstruction, allowing for optimized image generation without manual input.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of time

If measurement sequences are optimized and modified to reduce measurement time, then measurement time is reduced, but resolution and image quality deteriorate due to Nyquist theorem limitations

Engineering Contradiction:
Improvemeasurement timeVSAvoidimage resolution
Core Design Contradiction:
Loss of timeVSManufacturing precision

Solution Approach 1:

The patent applies preliminary action by performing a phase correction measurement before the actual image acquisition. This preliminary measurement captures phase information that is then used to correct the undersampled data, allowing for faster acquisition without sacrificing image quality. The phase correction data is prepared in advance to compensate for the effects of reduced sampling.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent introduces an intermediary element - the regularization parameter determined from phase correction measurements - that mediates between the conflicting requirements of fast acquisition and high resolution. This parameter acts as a bridge, allowing the reconstruction algorithm to recover high-resolution information from undersampled data by using the phase correction measurements as intermediate information.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Loss of time

If parallel imaging or compressed sensing is used to accelerate acquisition, then measurement time is reduced, but manual parameter tuning is required which increases complexity

Engineering Contradiction:
Improvemeasurement timeVSAvoidparameter tuning complexity
Core Design Contradiction:
Loss of timeVSDevice complexity

Solution Approach 1:

The patent implements self-service by enabling the system to automatically determine the regularization parameter using the measured phase correction data. The system serves itself by using its own measurement data (phase correction signals) to automatically set the reconstruction parameters, eliminating the need for manual intervention or expert knowledge in parameter tuning.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The patent applies feedback by using the phase correction measurements as feedback information to automatically adjust the regularization parameter in the reconstruction process. The system measures the phase corrections, feeds this information back into the parameter determination, and uses it to optimize the reconstruction, creating a closed-loop system that adapts to the actual measurement conditions.

Inventive Principle:
Principle #23Feedback

3Loss of time

If undersampled data is acquired to reduce measurement time, then measurement time is reduced, but noise and artifacts increase

Engineering Contradiction:
Improvemeasurement timeVSAvoidnoise and artifacts
Core Design Contradiction:
Loss of timeVSObject-affected harmful factors

Solution Approach 1:

The patent converts the harm of undersampling (which causes noise and artifacts) into a benefit by using the phase correction measurements to inform the reconstruction process. The regularization parameter, determined from phase correction data, helps the algorithm distinguish between actual image features and sampling-induced noise, effectively converting the problematic undersampled data into reconstructible information.

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

Data Source

PatentEP3564695B1Method for creating a magnetic resonance image data set, computer program product, data carrier and magnetic resonance system
Publication Date: 2025.06.25 SIEMENS HEALTHINEERS AG
  • EP3564695B1 patent drawingFigure 1
  • EP3564695B1 patent drawingFigure 2
  • EP3564695B1 patent drawingFigure 3~4

AI summary

The invention relates to a method for generating a magnetic resonance image dataset comprising the steps of: providing a raw dataset, wherein the raw dataset has been acquired with temporal and/or spatial undersampling, automated determination of a regularization parameter (λ), and generation of an image dataset from the raw dataset using the regularization parameter (λ) in a compressed sensing method. The invention further relates to a computer program product. The invention further relates to a data storage medium. The invention further relates to a magnetic resonance imaging system with which the aforementioned method can be carried out.