MRI Gradient Pre-emphasis Correction via Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Magnetic resonance imaging (MRI) systems face deviations from target magnetic field gradients due to factors like eddy currents, timing errors, amplifier errors, mechanical vibrations, and thermal fluctuations, leading to image artifacts and requiring complex correction methods that are not suitable for everyday clinical use.
Innovation Solution
A method and apparatus using a trained machine-learning function to reduce and correct deviations from target magnetic field gradients by providing input data about the target gradient, which is used to create output data for correcting amplifier input signals, thereby pre-emphasizing the magnetic field gradient to align it with the target gradient, and optionally applying post-processing corrections.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If complex correction methods are used to address gradient deviations, then image quality improves, but device complexity and processing time increase
Solution Approach 1:
The patent applies parameter changes by modifying the gradient waveform parameters (amplitude, duration, shape) through pre-emphasis filtering. The correction method adjusts the gradient pulse parameters in the time domain to compensate for expected deviations, transforming the gradient system's impulse response function into a corrected waveform that accounts for eddy currents and system dynamics without requiring complex post-processing algorithms
Solution Approach 2:
The patent implements preliminary action by performing gradient pre-emphasis before the actual MRI data acquisition. The system calculates the gradient system's impulse response function beforehand and uses it to pre-distort the gradient waveforms, so that the actual gradient applied during imaging already compensates for expected deviations. This eliminates the need for complex real-time correction during scanning
2Manufacturing precision
If gradient pre-emphasis is applied to correct deviations, then gradient accuracy improves, but system complexity increases
Solution Approach 1:
The patent employs feedback by measuring the actual gradient system response using a field camera or probe during a calibration phase, then using this measured impulse response function to calculate the optimal pre-emphasis filter. The system continuously refines the gradient correction based on measured performance, creating a closed-loop system that adapts to specific hardware characteristics without requiring complex real-time adjustments
Solution Approach 2:
The patent replaces complex mechanical adjustment systems with computational methods. Instead of using complex mechanical gradient adjustment mechanisms or hardware modifications, the system uses digital signal processing and mathematical modeling (impulse response function analysis) to achieve gradient precision, substituting mechanical complexity with algorithmic solutions
3Ease of operation
If gradient deviations are not corrected, then system operation remains simple, but image artifacts increase
Solution Approach 1:
The patent implements self-service by enabling the gradient system to automatically correct its own deviations through pre-emphasis filtering based on its measured impulse response function. The system performs self-diagnosis and self-correction without requiring external intervention or complex additional hardware, making the correction process as simple as loading and applying a pre-calculated filter during normal operation
Data Source
AI summary
In a method for reducing and/or correcting deviations from a target gradient of a magnetic field gradient created by an MR system input data is provided for a trained function trained by a machine-learning algorithm, wherein the input data comprises information about the target gradient of the MR system. The trained function further creates output data with the aid of the input data. The deviations from the target gradient of the magnetic field gradient created by the MR system are reduced and/or corrected with the aid of the output data created.