MRI Reconstruction Using Joint K-Space and Image-Space Calibration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing magnetic resonance imaging (MRI) methods for multi-frame data acquisition are limited by long scanning times and sensitivity to motion, and current parallel imaging techniques lack self-calibration, leading to artifacts and limited speed-up effects.
Innovation Solution
A joint k-space and image-space reconstruction imaging method (KIPI) that uses a self-calibration parallel imaging technique like GRAPPA to reconstruct undersampled frames, generating an optimized sensitivity map and applying a correction factor to suppress artifacts, allowing higher acceleration factors without additional sensitivity maps.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional parallel imaging methods (GRAPPA/SENSE) are used for multi-frame imaging, then reconstruction can be performed, but scanning time remains long and motion sensitivity increases
Solution Approach 1:
The patent performs preliminary action by acquiring a reference frame with full k-space data before undersampling subsequent frames. This reference frame is used to pre-calculate coil sensitivity maps and calibration weights, which are then applied to accelerate reconstruction of multiple subsequent frames without requiring additional calibration scans for each frame, thereby reducing total acquisition time
Solution Approach 2:
The patent makes the reference frame serve multiple functions: it provides calibration data for GRAPPA reconstruction, generates coil sensitivity maps for SENSE reconstruction, and enables artifact correction for all subsequent undersampled frames. This multi-functional use of a single reference frame eliminates the need for repeated calibration acquisitions across multiple frames
2Measurement precision
If GRAPPA method is used with low acceleration factor, then accurate images with high robustness are produced, but speed-up effect is limited
Solution Approach 1:
The patent dynamically adapts the reconstruction approach based on acceleration factor requirements. For low acceleration factors, traditional GRAPPA is used maintaining high accuracy. For high acceleration factors, the system switches to using pre-calculated sensitivity maps from the reference frame, enabling faster reconstruction while maintaining acceptable accuracy through the artifact correction mechanism
3Productivity
If GRAPPA method is used with high acceleration factor, then faster reconstruction is achieved, but artifacts increase and accuracy decreases
Solution Approach 1:
The patent introduces an intermediary artifact correction step that uses the reference frame as a mediator. The reference frame, acquired with full k-space data, serves as an intermediary structure to generate correction weights that are applied to the undersampled frames. This intermediary correction process removes aliasing artifacts and improves accuracy in high acceleration factor reconstructions without requiring slower traditional calibration for each frame
4Reliability
If multiple ACS frames are collected for each image frame, then self-calibration is achieved, but scanning time increases significantly
Solution Approach 1:
The patent implements self-service by making the first undersampled frame serve its own calibration needs. This frame contains embedded ACS data that is used to calculate GRAPPA weights and generate sensitivity maps specifically for itself and subsequent frames. This self-calibration approach eliminates the need for separate calibration scans before each undersampled frame, reducing total acquisition time while maintaining calibration accuracy
5Productivity
If only one ACS frame is collected for all undersampled frames, then scanning time is reduced, but unknown artifacts are introduced
Solution Approach 1:
The patent implements feedback by using the reconstructed image from the first undersampled frame to generate artifact correction weights that are then applied back to correct artifacts in all subsequent undersampled frames. This feedback loop allows the system to identify and correct artifacts based on the actual reconstructed data rather than relying solely on pre-calculated calibration parameters
Data Source
AI summary
A joint k-space and image-space reconstruction imaging method and a device, relating to the field of magnetic resonance imaging. The method comprises: first, using the k-space parallel imaging method to reconstruct under-sampled image frames having auto-calibration signals of a subject obtained from multi-frame imaging required for the measurement of magnetic resonance parameters; and then performing reconstruction in image space for other under-sampled image frame data without auto-calibration signals. In the method, an accurate sensitivity map required for the SENSE method is generated from under-sampled data, without the need of additional acquisition. By means of the present method, images that are almost identical to those obtained by the traditional GRAPPA method and without obvious artifacts can be obtained at a speed which is four times faster than that by the traditional GRAPPA method. The present method is especially suitable for accelerating three-dimensional (3D) multi-frame magnetic resonance imaging.