Coil Mixing Error Matrix Deep Learning Motion Assessment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Magnetic resonance (MR) image acquisition is hindered by motion artifacts, leading to reduced image quality and potential misinterpretation, as existing methods either require additional equipment or are performed retrospectively, limiting their effectiveness.
Innovation Solution
A method and system that utilize a neural network to prospectively identify motion by calculating coil mixing matrices and error matrices from reference data, allowing for real-time motion assessment during MR image acquisition without additional hardware, enabling adaptive acquisition parameters and retrospective correction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If additional equipment is used to correct motion artifacts, then image quality is improved, but device complexity increases
Solution Approach 1:
The patent uses the existing MR imaging system's own coil signals to detect and correct motion artifacts. The coil mixing matrices are calculated from the MR data itself, allowing the system to self-diagnose and self-correct motion issues without requiring external motion tracking equipment or additional hardware.
Solution Approach 2:
The patent extracts motion information from the coil signal variations in the MR data. By analyzing how the coil mixing matrices change between reference and current data, the system isolates motion effects from the imaging signals, enabling motion correction using only the existing MR imaging system.
2Manufacturing precision
If retrospective motion correction is performed, then image quality is improved, but loss of time occurs
Solution Approach 1:
The patent implements feedback by continuously monitoring coil signal variations during the MR acquisition and providing real-time motion assessment. The neural network processes coil mixing matrices as they are generated, enabling prospective motion detection that can trigger immediate corrective actions or warnings during the imaging procedure.
Solution Approach 2:
The patent calculates coil mixing matrices from a motion-free reference dataset before the actual imaging acquisition. This preliminary calculation establishes a baseline that enables real-time motion detection during scanning, allowing motion artifacts to be identified and corrected during rather than after the procedure.
3Manufacturing precision
If motion assessment is performed during acquisition, then image quality is improved, but use of energy increases
Solution Approach 1:
The patent applies motion assessment selectively to specific coil mixing matrix calculations rather than processing all MR data in real-time. The neural network is trained offline on partial datasets, and during acquisition, only the essential coil mixing matrix comparisons are performed, reducing real-time computational energy requirements while maintaining effective motion detection.
Data Source
AI summary
Systems and Methods that identify the effect of motion during a medical imaging procedure. A neural network is trained to translate motion induced deviations of a coil-mixing matrix relative to a reference acquisition into a motion score. This score can be used for the prospective detection of the most corrupted echo trains for removal or triggering a replacement by reacquisition.