MRI Motion Correction Guided by Neural Trajectory Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Magnetic Resonance Imaging (MRI) is challenged by subject motion during k-space data acquisition, leading to blurring and image artifacts due to the numerical intensity and risk of settling into local minima in traditional optimization methods for motion correction.
Innovation Solution
A motion estimating neural network is used to modify the optimization problem's cost function with a trajectory probability term, guiding the search towards likely solutions, reducing computational burden and improving convergence by emphasizing more probable motion trajectories.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional numerical optimization methods are used for motion correction, then motion correction can be performed, but computational load is very high and risk of settling into local minima increases
Solution Approach 1:
A neural network is introduced as an intermediary component that predicts motion trajectories from k-space data. This neural network acts as a mediator between the raw data and the optimization process, providing informed initial motion estimates that guide the subsequent numerical optimization, thereby reducing both computational load and the risk of local minima
Solution Approach 2:
The neural network performs preliminary motion trajectory prediction before the main optimization process. By providing pre-computed motion estimates based on learned patterns from training data, the system prepares informed initial conditions that accelerate convergence and improve reliability of the final motion correction
2Reliability
If traditional numerical optimization methods are used for motion correction, then motion correction can be performed, but computational load is very high
Solution Approach 1:
The neural network serves as an energy-efficient intermediary that performs preliminary motion estimation using learned patterns rather than exhaustive numerical search. This reduces the computational energy required for the subsequent optimization process while maintaining motion correction accuracy
Solution Approach 2:
By performing preliminary motion trajectory prediction using the trained neural network, the system avoids the high energy consumption of exhaustive numerical optimization. The pre-computed motion estimates enable the optimization to converge faster with significantly reduced computational energy requirements
3Measurement precision
If exhaustive numerical search is performed for motion trajectories, then accurate motion correction may be achieved, but computation time increases significantly
Solution Approach 1:
The neural network acts as a time-efficient mediator that provides accurate motion trajectory predictions based on learned patterns from training data. This intermediary approach replaces exhaustive numerical search with a faster inference process that maintains trajectory accuracy while dramatically reducing computation time
Solution Approach 2:
The system performs preliminary motion trajectory estimation using the trained neural network before conducting detailed optimization. This preliminary action provides accurate initial motion estimates that guide the optimization process, achieving high measurement precision with significantly reduced computation time compared to exhaustive search methods
Data Source
AI summary
Described herein is a medical system (100, 300) comprising a memory (110) storing machine executable instructions (120) and a motion estimating neural network (122, 700, 800, 900, 1000) configured for outputting trajectory data (130) in response to receiving a trial motion trajectory (128) as input. The execution of the machine executable instructions causes a computational system (104) to: receive (200) measured k-space data (124) descriptive of a subject (318); perform (202) motion estimation of the subject between the sequence of discrete acquisitions by solving an optimization problem to determine a calculated motion trajectory of the subject in the predefined coordinate system, wherein the optimization problem is modified using the trajectory data; and reconstruct (204) a final motion corrected magnetic resonance image (136) from the measured k-space data and the calculated motion trajectory in the predefined coordinate system.