Joint MRI Reconstruction and Segmentation Using Deep Neural Networks

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current MRI reconstruction and segmentation techniques are time-consuming, require manual intervention, and struggle with motion artifacts, especially in cardiac MRI, due to the need for fully-sampled data and breath-hold constraints, which limits their clinical application and accuracy.

Innovation Solution

The proposed system employs a joint reconstruction and segmentation process using a channel-wise attention network for static reconstruction and a motion-guided network for dynamic reconstruction, leveraging deep neural networks to process sparse k-space MRI data, thereby reducing scan and reconstruction time while improving image quality and segmentation accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If fully-sampled MRI scan is used, then image quality is improved, but scan time increases significantly

Engineering Contradiction:

Improveimage qualityVSAvoidscan time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The patent uses partial sampling of k-space data (acquiring only a subset of the full data) combined with deep learning reconstruction to achieve acceptable image quality without the need for complete sampling, thereby reducing scan time while maintaining diagnostic utility

Inventive Principle:

Principle #16Partial or excessive action

Solution Approach 2:

The patent replaces traditional iterative reconstruction algorithms with deep neural network-based reconstruction that can rapidly process undersampled k-space data, substituting computational mechanics with learned patterns to achieve fast and accurate image recovery

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Reliability

If breath-hold constraint is enforced, then motion artifacts are reduced, but patient comfort and compliance deteriorate

Engineering Contradiction:

Improveimage accuracyVSAvoidpatient compliance

Core Design Contradiction:

ReliabilityVSEase of operation

Solution Approach 1:

The patent converts the harmful effect of motion during scanning into a beneficial outcome by using motion information itself as input to the deep learning model, which learns to compensate for and correct motion artifacts rather than requiring their prevention through breath-holding

Inventive Principle:

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

Solution Approach 2:

The patent transitions from static image acquisition requiring breath-holds to dynamic image acquisition that captures motion and uses it for reconstruction, allowing patients to breathe naturally while the system adapts to the actual motion patterns

Inventive Principle:

Principle #15Dynamics

3Measurement precision

If manual intervention is used for segmentation, then segmentation accuracy is improved, but processing time and complexity increase

Engineering Contradiction:

Improvesegmentation accuracyVSAvoidprocessing speed

Core Design Contradiction:

Measurement precisionVSProductivity

Solution Approach 1:

The patent implements self-service by having the deep learning model perform segmentation automatically without human intervention, where the system itself handles the complex task of identifying and segmenting cardiac structures from the reconstructed images

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent creates a multi-functional deep learning system that performs both reconstruction and segmentation tasks, allowing a single model to handle multiple processing steps that traditionally required separate manual operations

Inventive Principle:

Principle #6Universality (Multi-functionality)

4Measurement precision

If traditional reconstruction methods are used, then reconstruction accuracy is maintained, but reconstruction time increases

Engineering Contradiction:

Improvereconstruction accuracyVSAvoidreconstruction time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The patent replaces traditional iterative mathematical reconstruction algorithms with deep neural network-based reconstruction that uses learned patterns from training data to rapidly recover images from undersampled k-space, achieving both speed and accuracy

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentUS11645791B2Systems and methods for joint reconstruction and segmentation of organs from magnetic resonance imaging data

Publication Date: 2023.05.09 RUTGERS THE STATE UNIV

AI summary

Systems and methods for joint reconstruction and segmentation of organs from magnetic resonance imaging (MRI) data are provided. Sparse MRI data is received at a computer system, which jointly processes the MRI data using a plurality of reconstruction and segmentation processes. The MRI data is processed using a joint reconstruction and segmentation process to identify an organ from the MRI data. Additionally, the MRI data is processed using a channel-wise attention network to perform static reconstruction of the organ from the MRI data. Further, the MRI data can is processed using a motion-guided network to perform dynamic reconstruction of the organ from the MRI data. The joint processing allows for rapid static and dynamic reconstruction and segmentation of organs from sparse MRI data, with particular advantage in clinical settings.