Neural Network Motion Artifact Detection in MRI K-Space Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Motion artifacts in MRI scans due to patient movement cause image distortions and degradations, leading to clinical diagnosis challenges and increased radiology workflow and costs, as existing technologies fail to effectively detect and correct these issues in real-time.
Innovation Solution
A method involving the training of a neural network to detect motion artifacts in k-space data using simulated data generated by a forward model, which applies varying degrees of motion to MRI images, allowing for the classification of artifact presence and severity, enabling immediate corrective action during scans.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If real-time motion artifact detection is implemented during MRI scans, then image quality and diagnostic accuracy are improved, but device complexity and computational requirements increase
Solution Approach 1:
The neural network is trained offline using simulated motion-corrupted k-space data generated by a forward model. This preliminary training phase allows the system to learn motion artifact patterns without adding real-time computational complexity to the actual scanning process. The pre-trained model is then deployed for rapid inference during scans.
Solution Approach 2:
The patent uses simulated k-space data that copies realistic motion artifacts through a forward model, rather than requiring large volumes of actual motion-corrupted clinical data. This copying approach enables effective training while reducing the complexity of data collection and annotation processes.
2Productivity
If motion artifacts are detected during scanning, then patient recall is reduced and workflow efficiency improves, but processing time and computational resources increase
Solution Approach 1:
The patent extracts only the essential features needed for motion artifact detection from the full k-space data, rather than processing the entire dataset. The neural network is designed to identify specific motion-related patterns in k-space, enabling rapid detection without analyzing all image data in detail.
Solution Approach 2:
The system performs partial processing by focusing computational resources on detecting motion artifacts rather than fully reconstructing and analyzing the complete image. This selective approach provides sufficient information for workflow decisions without the computational overhead of complete image processing.
3Ease of manufacture
If simulated training data is used to train the neural network, then data acquisition and labeling effort is reduced, but training data realism and model generalization may be affected
Solution Approach 1:
The patent converts the limitation of not having access to large volumes of actual motion-corrupted clinical data into a benefit by using a forward model to generate realistic simulated training data. The forward model intentionally introduces motion artifacts into synthetic k-space data, creating a reliable training dataset that would otherwise be difficult to obtain.
Solution Approach 2:
The forward model allows systematic variation of motion parameters (amplitude, frequency, direction) to generate diverse training examples. By changing these parameters, the system creates a comprehensive training dataset that improves model generalization without requiring manual data collection for each scenario.
Data Source
AI summary
Motion artifacts are detected from raw k-space data acquired with a magnetic resonance imaging (“MM”) system. A machine learning model is trained on a training dataset that includes motion-simulated k-space data. The motion-simulated k-space data may be generated by inputting magnetic resonance images to a forward model to convert the images to k-space data while adding motion based on motion parameters. The severity of the simulated motion can be varied, and features of motion artifacts extracted by preprocessing the motion-simulated k-space data. In deployment, the trained machine learning model may be used to detect the presence and/or severity of motion artifacts in k-space data while a subject is being scanned with an MRI scanner.