Hidden Markov Model Motion Estimation in MRI Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current MRI techniques face challenges in accurately correcting motion artifacts, particularly due to physiological movements like cardiac pulsations and respiration, which can lead to image degradation, and existing retrospective motion correction techniques often introduce blurring or fail to account for intra-scan motion effectively.
Innovation Solution
The implementation of a hidden Markov model for dynamic motion estimation, using recursive and non-recursive Bayesian estimation techniques, such as the Extended Kalman Filter, to process MRI data and adjust imaging gradients and RF signals in real-time, thereby reducing motion-induced artifacts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If retrospective motion correction techniques are used to correct motion artifacts after data collection, then motion artifacts can be reduced, but image blurring is introduced and spin history effects cannot be corrected
Solution Approach 1:
The patent applies preliminary action by performing motion estimation and correction during the data acquisition process itself, rather than after collection. The Hidden Markov Model continuously estimates motion parameters as data is being acquired, allowing real-time correction of k-space data before it is fully collected, thereby avoiding the blurring and spin history effects that plague retrospective correction methods
Solution Approach 2:
The patent implements feedback by using the Hidden Markov Model to continuously estimate motion parameters from acquired data and feed this information back into the imaging process. The estimated motion parameters are used to update the imaging trajectory in real-time, creating a closed-loop system that actively compensates for motion during acquisition rather than correcting it afterward
2Measurement precision
If multiple navigator echoes are repeated to obtain more accurate motion estimates, then measurement precision improves, but scanning time increases
Solution Approach 1:
The patent merges the motion estimation function with the main imaging data acquisition process. Instead of using separate navigator echoes that repeat after the main scan, the Hidden Markov Model processes the k-space data being acquired during the main imaging sequence itself, combining two functions (imaging and motion tracking) into a single integrated process that does not increase scan time
Solution Approach 2:
The patent applies self-service by enabling the main imaging data to serve dual purposes: both for image formation and for motion estimation. The Hidden Markov Model extracts motion information from the k-space data being acquired for imaging, allowing the imaging process itself to provide the motion correction information without requiring additional dedicated motion tracking acquisitions
3Object-affected harmful factors
If physical restraints are used to immobilize the subject during scanning, then motion artifacts are reduced, but subject comfort decreases and restraints cannot fully prevent motion
Solution Approach 1:
The patent replaces the mechanical restraint system with a computational approach. Instead of using physical devices to constrain subject movement, the Hidden Markov Model computationally estimates motion from the acquired k-space data and corrects for it during the imaging process, eliminating the need for uncomfortable physical restraints while still achieving motion artifact reduction
Data Source
AI summary
Systems, apparatus and methods that use a hidden Markov model to estimate motion of a sample under measurement and to reduce a motion-induced effect in the measurement. MRI systems and other sample measurement systems can be implemented based on motion sensing using the hidden Markov model. Recursive and non-recursive estimation processes are described.