MRI Self-Calibration Using Deep Learning to Reduce Reconstruction Time
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current self-calibration and reconstruction methods for wave-encoded single-shot fast spin echo (SSFSE) MRI are computationally expensive and time-consuming, leading to delays in image acquisition and potential degradation of image quality due to respiratory and cardiac motion, especially in abdominal applications.
Innovation Solution
The use of deep neural networks for data-driven self-calibration and reconstruction of wave-encoded SSFSE, which estimates systematic waveform errors and computes the magnetic-field-gradient-encoding point-spread function to accelerate the calibration and reconstruction process, reducing computation time and improving image quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional self-calibration and reconstruction methods are used for wave-encoded SSFSE, then image quality can be maintained, but computation time increases significantly
Solution Approach 1:
The patent pre-calculates and stores calibration parameters (point spread function, sensitivity profiles) during a preliminary calibration scan. These pre-computed parameters are then reused during the actual imaging process, eliminating the need for time-consuming iterative optimization during reconstruction. This preliminary action resolves the contradiction by maintaining image quality through accurate calibration while drastically reducing computation time during clinical imaging.
Solution Approach 2:
The patent applies partial action by using a reduced set of calibration measurements compared to full calibration approaches. Instead of calibrating all possible parameters across the entire field of view, the method uses a subset of measurements that are sufficient for accurate reconstruction. This partial calibration approach maintains adequate image quality while significantly reducing the computational burden and scan time.
2Measurement precision
If iterative optimization is performed for self-calibration of point-spread function, then calibration accuracy improves, but computation cost increases
Solution Approach 1:
The patent uses a measured point spread function from a calibration scan as a template or copy that is then applied to the actual imaging data. Instead of performing iterative optimization for each imaging dataset, the pre-measured PSF copy is reused. This copying approach maintains calibration accuracy while eliminating the repeated computational cost of iterative optimization, directly resolving the contradiction between calibration accuracy and computation cost.
3Manufacturing precision
If parallel imaging and compressed sensing reconstruction is used, then image sharpness improves, but reconstruction time increases
Solution Approach 1:
The patent pre-computes and stores the point spread function and coil sensitivity profiles during a preliminary calibration phase. These pre-computed parameters enable direct application of parallel imaging and compressed sensing algorithms during reconstruction without requiring iterative optimization. This preliminary action maintains image sharpness through accurate point spread function modeling while dramatically reducing reconstruction time by eliminating repeated computational iterations.
Data Source
AI summary
A method for magnetic resonance imaging (MRI) includes steps of acquiring by an MRI scanner undersampled magnetic-field-gradient-encoded k-space data; performing a self-calibration of a magnetic-field-gradient-encoding point-spread function using a first neural network to estimate systematic waveform errors from the k-space data, and computing the magnetic-field-gradient-encoding point-spread function from the systematic waveform errors; reconstructing an image using a second neural network from the magnetic-field-gradient-encoding point-spread function and the k-space data.