MRI Time-Series Reconstruction From Undersampled K-Space Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Dynamic magnetic resonance imaging requires long scan times due to the need for acquiring all k-space frame data, leading to temporally discontinuous magnetic resonance image frames that cannot accurately represent temporal changes, and existing reconstruction methods suffer from artifacts and reduced accuracy when undersampling is employed.
Innovation Solution
An image processing apparatus and method that utilizes a neural network-based sensitivity map calculation and regularization processes to reconstruct magnetic resonance image time series data from undersampled k-space data, incorporating ACS extraction, grouping, and data consistency techniques to improve accuracy and reduce artifacts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If all k-space frame data is acquired to ensure complete information, then manufacturing precision is improved, but loss of time increases
Solution Approach 1:
The patent applies undersampling by acquiring only a subset of k-space frame data rather than all data points. This partial action reduces scan time while the iterative reconstruction algorithm compensates for the missing data to maintain acceptable image quality, directly resolving the contradiction between complete data acquisition and time efficiency
2Loss of time
If undersampling is used to reduce scan time, then loss of time is reduced, but measurement precision deteriorates
Solution Approach 1:
The patent employs dynamic temporal regularization that adapts to the temporal characteristics of the imaged object. The regularization strength and parameters are adjusted dynamically based on the observed temporal changes, allowing accurate representation of temporal dynamics even with undersampled data, thus resolving the contradiction between reduced scan time and temporal accuracy
3Productivity
If undersampling is employed to accelerate acquisition, then productivity is improved, but object-generated harmful factors increase
Solution Approach 1:
The patent implements iterative reconstruction with feedback mechanisms where the reconstructed images are continuously refined by comparing with the acquired undersampled data and applying regularization constraints. This feedback loop suppresses artifact generation while maintaining acquisition speed, resolving the contradiction between productivity and artifact reduction
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The method achieves higher stability and accuracy in magnetic resonance image reconstruction, reducing artifacts and improving the structural similarity and peak signal-to-noise ratio while maintaining accuracy even with undersampling, thereby enhancing the representation of temporal changes.
Implementation Method 1
calculates sensitivity maps of a plurality of receiver coils via a neural network
Implementation Method 2
a magnetic resonance imaging technique using a magnetic resonance phenomenon in which a hydrogen nucleus placed in a static magnetic field resonates with a high-frequency magnetic field
Implementation Method 3
a pulse signal is transmitted to a subject (patient) in a magnetic field that is frequency-encoded and phase-encoded, and echo signals due to specific nuclear magnetic resonance are received from a plurality of receiver coils
Data Source
AI summary
An image processing apparatus according to an embodiment includes processing circuitry. The processing circuitry acquires k-space data, classifies ACS data in the k-space data corresponding to each of a plurality of frames into a plurality of groups, generates a sensitivity map, performs an image space regularization process, and performs a data consistency process.