Neural Network MRI Artifact Detection and Pulse Sequence Modification

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Magnetic Resonance Imaging (MRI) systems face challenges with artifacts due to subject motion, metallic implants, and field inhomogeneities, leading to image degradation and the need for extensive re-scanning.

Innovation Solution

A neural network-based MRI artifact detection module identifies artifacts and suggests pulse sequence command changes to improve image quality, using convolutional neural networks trained with artifact-corrupted images and corresponding labels to provide probability scores for image improvement.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If traditional MRI scanning is performed without artifact detection, then the imaging workflow is simple, but artifacts cause image degradation and require extensive re-scanning

Engineering Contradiction:

Improveimaging workflow efficiencyVSAvoidimage quality

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The system performs preliminary artifact detection and classification on the acquired MRI image before final image production. The neural network analyzes the image for artifacts and generates classification results that guide subsequent pulse sequence modifications, preventing the need for complete re-scanning and improving workflow efficiency

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system implements a feedback loop where artifact detection results inform pulse sequence command modifications. The classification of detected artifacts feeds back into adjusting imaging parameters, which then improves the quality of subsequent images without requiring extensive re-scanning

Inventive Principle:

Principle #23Feedback

2Reliability

If artifact detection and correction is implemented, then image quality improves and re-scanning is reduced, but the system complexity increases

Engineering Contradiction:

Improveimage qualityVSAvoidsystem complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The patent introduces an intermediary neural network-based artifact detection module that sits between the MRI scanner and the reconstruction software. This intermediary automatically classifies artifacts and suggests pulse sequence modifications, adding intelligence without requiring complex manual intervention or complete system redesign

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The system performs self-diagnosis by automatically detecting and classifying artifacts in the acquired images. The neural network independently analyzes image data, identifies artifact types, and generates recommendations for pulse sequence adjustments without requiring external expert intervention, thereby managing complexity internally

Inventive Principle:

Principle #25Self-service

3Loss of time

If manual artifact identification and correction is performed, then the system remains simple, but time is lost due to extensive re-scanning

Engineering Contradiction:

Improvere-scanning timeVSAvoidartifact detection automation

Core Design Contradiction:

Loss of timeVSExtent of automation

Solution Approach 1:

The patent replaces the manual mechanical process of artifact identification and correction with an automated neural network-based system. The neural network automatically analyzes images, classifies artifacts, and suggests pulse sequence modifications, eliminating the need for manual review and reducing re-scanning time

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system automatically modifies pulse sequence parameters based on neural network classification results. By changing imaging parameters such as echo train length, echo spacing, or bandwidth according to the detected artifact type, the system reduces artifacts in subsequent scans without manual intervention, thereby reducing time loss

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS11333732B2Automatic artifact detection and pulse sequence modification in magnetic resonance imaging

Publication Date: 2022.05.17 KONINKLIJKE PHILIPS NV

AI summary

The invention provides for a magnetic resonance imaging system (100, 300). The execution of machine executable instructions causes a processor (130) controlling the magnetic resonance imaging system to control (200) the magnetic resonance imaging system to acquire the magnetic resonance imaging data (144) using pulse sequence commands (142) and reconstruct (202) a magnetic resonance image (148). Execution of the machine executable instructions causes the processor to receive (204) a list of suggested pulse sequence command changes (152) by inputting the magnetic resonance image and image metadata (150) into an MRI artifact detection module (146, 146′, 146″). The MRI artifact detection module comprises at least one neural network, which has been trained using images from failed magnetic resonance imaging protocols and/or magnetic resonance data extracted from the magnetic resonance imaging protocols labeled as failed accessed from a log file (312) which logs the execution of previous magnetic resonance imaging protocols. Execution of the machine executable instructions further causes the processor to receive (206) a selection of a chosen pulse sequence command change (158) from the list of suggested pulse sequence command changes. Execution of the machine executable instructions further causes the processor to modify (208) the pulse sequence commands using the chosen pulse sequence command change.