Compressed Sensing MRI Outlier Data Exclusion

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Compressed sensing magnetic resonance imaging (MRI) techniques face challenges in accurately reconstructing images due to motion artifacts and data inconsistencies, leading to suboptimal image quality and increased computational intensity.

Innovation Solution

The method involves reconstructing an intermediate MRI image using compressed sensing protocols, calculating predicted k-space data, and identifying outlier data portions by comparing measured and predicted data, which are then excluded from the reconstruction process to improve image quality and reduce artifacts.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If compressed sensing MRI is used to reduce scan time and computational load, then productivity is improved, but image quality deteriorates due to motion artifacts and data inconsistencies

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

Solution Approach 1:

The patent applies preliminary action by performing an initial reconstruction to create a preliminary image before the final reconstruction. This preliminary image is then used to identify and exclude outlier data portions, which improves the quality of the final reconstruction without requiring additional scan time. The process prepares the data in advance by flagging problematic portions based on the preliminary image quality metrics.

Inventive Principle:
Principle #10Preliminary action

2Manufacturing precision

If more k-space data is collected to improve image quality, then manufacturing precision is improved, but the quantity of data increases leading to higher computational complexity

Engineering Contradiction:
Improveimage qualityVSAvoidcomputational complexity
Core Design Contradiction:
Manufacturing precisionVSDevice complexity

Solution Approach 1:

The patent extracts and removes outlier data portions from the k-space dataset before performing the final reconstruction. By identifying data portions that do not conform to expected quality criteria (using metrics like signal intensity thresholds or coherence measures) and excluding them from the reconstruction process, the method reduces computational complexity while maintaining or improving image quality. This extraction of problematic data prevents it from degrading the final image and reduces the computational burden of processing inconsistent data.

Inventive Principle:
Principle #2Taking out (Extraction)

3Manufacturing precision

If iterative reconstruction is performed multiple times to improve image quality, then manufacturing precision is improved, but loss of time increases due to repeated computations

Engineering Contradiction:
Improveimage qualityVSAvoidreconstruction time
Core Design Contradiction:
Manufacturing precisionVSLoss of time

Solution Approach 1:

The patent performs a preliminary reconstruction to generate a preliminary image that is then used to identify outlier data portions. This preliminary action allows the method to exclude problematic data before the final reconstruction, reducing the number of iterations needed and thereby reducing total reconstruction time while maintaining image quality.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

By extracting and removing outlier data portions identified from the preliminary image, the patent reduces the computational burden of the final reconstruction. The exclusion of inconsistent data portions allows for faster convergence in the iterative reconstruction process, reducing the time required to achieve high-quality images.

Inventive Principle:
Principle #2Taking out (Extraction)

4Manufacturing precision

If outlier data portions are excluded from reconstruction, then image quality is improved by reducing motion artifacts, but loss of information occurs by removing potentially useful data

Engineering Contradiction:
Improveimage qualityVSAvoiddata completeness
Core Design Contradiction:
Manufacturing precisionVSLoss of information

Solution Approach 1:

The patent applies local quality by selectively excluding only the outlier data portions that exhibit poor quality characteristics (such as high noise levels, inconsistency with neighboring data, or failure to meet signal intensity thresholds) while retaining and using the majority of the k-space data for reconstruction. This localized exclusion approach maintains image quality by removing only the problematic portions rather than discarding large amounts of potentially useful data.

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP3500869B1Corrected compressed sensing magnetic resonance imaging
Publication Date: 2021.02.24 KONINKLIJKE PHILIPS NV
  • EP3500869B1 patent drawingFigure 1
  • EP3500869B1 patent drawingFigure 2
  • EP3500869B1 patent drawingFigure 3

AI summary

The invention provides for a magnetic resonance imaging system (100) comprising: a memory (150) for storing machine executable instructions (160) and for storing pulse sequence commands (162) to acquire the measured magnetic resonance data according to a compressed sensing magnetic resonance imaging protocol; and a processor (144) for controlling the magnetic resonance imaging system. Execution of the machine executable instructions cause the processor to: control (200) the magnetic resonance imaging system with the pulse sequence commands to acquire the measured magnetic resonance data, wherein the measured magnetic resonance data is acquired as measured data portions (164), wherein each of the measured data portions is acquired during a time period; reconstruct (202) an intermediate magnetic resonance image (168) using the measured magnetic resonance data according to the compressed sensing magnetic resonance imaging protocol; calculate (204) a predicted data portion (170) for each of the measured data portions using the intermediate magnetic resonance image; calculate (206) a residual (172) for each of the measured data portions using the predicted data portion; identify (208) one or more of the measured data portions as outlier data portions (176) if the residual is above a predetermined threshold; and reconstruct (210) a corrected magnetic resonance image (178) using the measured magnetic resonance data according to the compressed sensing magnetic resonance imaging protocol, and wherein the one or more outlier data portions are excluded from the reconstruction of the corrected magnetic resonance image.