3D MRI Coil Compression via Slice-Aligned Hybrid Space Matrices
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current 3D magnetic resonance imaging (MRI) data acquisition techniques are limited by long scan times, leading to undesirable artifacts and high computation costs in reconstruction, especially with large phased arrays, due to inefficient coil compression methods that do not account for spatially varying coil sensitivities.
Innovation Solution
A slice-by-slice coil compression method is introduced, where compression matrices are carefully aligned to minimize the number of virtual coils and reduce reconstruction time, allowing for efficient data-based coil compression that maintains image quality by incorporating spatially varying coil sensitivities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If conventional coil compression methods are used for 3D MRI data acquisition, then the number of coils can be reduced, but reconstruction time becomes excessively long and computation cost increases
Solution Approach 1:
The patent divides the 3D k-space data into multiple 2D slices along the readout direction. Each slice is processed independently with its own compression matrix, allowing parallel computation and significantly reducing reconstruction time compared to processing the entire 3D dataset as a single unit.
Solution Approach 2:
The patent performs coil compression on each 2D slice before the final 3D reconstruction. By pre-compressing the data slice-by-slice using alignment matrices that account for spatially varying coil sensitivities, the computation burden is reduced in advance, making the subsequent reconstruction process much faster.
2Device complexity
If conventional coil compression methods are used, then device complexity is reduced, but manufacturing precision of image quality deteriorates due to ignoring spatially varying coil sensitivities
Solution Approach 1:
The patent applies different compression matrices to different slices along the readout direction, allowing each slice to be compressed according to its local coil sensitivity characteristics. This spatially varying approach preserves image quality by adapting to the local properties of the phased array coils rather than using a uniform compression approach.
Solution Approach 2:
The patent changes the compression parameters (compression matrices) as a function of position along the readout direction. By computing separate compression matrices for each slice based on the spatially varying coil sensitivities, the system maintains optimal compression performance across different regions of the 3D dataset.
3Measurement precision
If 3D Cartesian data acquisition is used, then signal-to-noise ratio and volume coverage are improved, but scan time becomes too long for clinical application
Solution Approach 1:
The patent applies compressed sensing techniques that allow random undersampling of k-space data. By acquiring only a subset of the required k-space samples (partial action) and using iterative reconstruction algorithms that exploit image sparsity, the system achieves full-resolution images with significantly reduced scan times while maintaining signal-to-noise ratio.
Data Source
AI summary
A three dimensional image, in a phased array magnetic resonance imaging (MRI) system is provided. Three dimensional k-space data within an auto calibration signal (ACS) region and outside the ACS region are acquired. The k-space data within the ACS region are converted into hybrid space ACS data. Compression matrices and alignment matrices of the compression matrices for the hybrid space ACS data are found along a readout direction. Alignment matrices are multiplied to the compression matrices to achieve the properly-aligned compression matrices along the readout direction. All k-space data are converted into hybrid space. The properly-aligned compression matrices are applied to the hybrid space data to provide compressed data with fewer channels. The compressed data are used to form a three dimensional image.