Self-Calibrated Parallel MRI Reconstruction via Integrated K-Space Sampling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing parallel MRI techniques require additional calibration data acquisition, which increases overhead time and reduces the acceleration factor, necessitating longer acquisition times and additional constraints on data acquisition.
Innovation Solution
Integrating calibration data acquisition with image data acquisition in a time-series k-space data set, allowing calibration data to be sampled across multiple time frames without reducing the acceleration factor, thereby reducing the overall scan time and maintaining high spatial resolution.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If separate calibration data acquisition is performed, then calibration accuracy is improved, but acquisition time increases and acceleration factor decreases
Solution Approach 1:
The patent combines calibration data acquisition with image data acquisition into a single integrated process. Calibration data and image data are acquired simultaneously during the same scan, eliminating the need for separate calibration scans. This merging approach maintains calibration accuracy while significantly reducing total acquisition time and preserving the acceleration factor.
Solution Approach 2:
The patent makes the acquisition process multi-functional by designing a unified data acquisition scheme that serves both calibration and imaging purposes. The same acquisition sequence generates both calibration data (from coil sensitivity profiles) and image data (from the imaged object), allowing one process to fulfill multiple functions without requiring additional time or resources.
2Measurement precision
If separate calibration data acquisition is performed, then calibration data quality is improved, but acceleration factor is reduced
Solution Approach 1:
The patent merges calibration data acquisition with image data acquisition, allowing both types of data to be collected during the same accelerated scan. This eliminates the sequential overhead that would reduce the acceleration factor, while the integrated processing maintains calibration data quality through proper signal separation and reconstruction algorithms.
Solution Approach 2:
The patent ensures continuous useful action by acquiring both calibration and image data continuously during the same scan without interruption or sequential delays. The acquisition process flows continuously, maximizing the acceleration factor while maintaining data quality through uninterrupted signal collection from all coils.
3Loss of time
If calibration data is acquired within the accelerated acquisition, then total scan time is reduced, but the acceleration factor must be reduced to accommodate calibration data
Solution Approach 1:
The patent merges calibration data and image data into a single acquisition stream, allowing both to be collected during the same accelerated scan at the full acceleration factor. The combined data are processed separately through reconstruction algorithms that distinguish between calibration signals and image signals, eliminating the need to reduce acceleration to accommodate calibration data.
Solution Approach 2:
The patent segments the acquired data into calibration data and image data components during processing, even though both were acquired simultaneously. This segmentation allows the full acceleration factor to be maintained during acquisition while still extracting high-quality calibration data for parallel imaging reconstruction.
Data Source
AI summary
A method for producing a time-series of images of a subject with a magnetic resonance imaging (MRI) system is provided. The MRI system is used to acquire a time-series undersampled k-space data set, in which a selected number of k-space data subsets in the time-series data set includes both image data and calibration data. Moreover, the calibration data in each of these selected number of k-space data subsets includes a portion of a desired total amount of calibration data. For example, each of these selected number of k-space data subsets include calibration data that is acquired by sampling a different partition of a calibration data sampling pattern. A time-series of images of the subject is then produced by reconstructing images of the subject from the acquired time-series of undersampled k-space data sets. These images are substantially free of undersampling artifacts.