Randomized k-space undersampling for MRI acquisition speed
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current magnetic resonance imaging (MRI) techniques fail to accelerate the acquisition time for two-dimensional volume segments or slices, which accounts for 90% of MR images, despite advancements in three-dimensional imaging.
Innovation Solution
A method involving the randomized determination and undersampling of k-space points using radial or spiral trajectories, with a pseudo-random distribution that decreases with distance from the center, allowing for efficient acquisition of MR data while maintaining image quality even at high undersampling rates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If k-space is sampled in its entirety using conventional methods, then complete MR data is acquired, but acquisition time is excessively long
Solution Approach 1:
The patent applies partial sampling by acquiring only a subset of k-space points rather than the complete k-space grid. Randomly selected k-space points are sampled with probability p, acquiring sufficient information for image reconstruction while significantly reducing the number of measurements needed, thus resolving the contradiction between data completeness and acquisition time
Solution Approach 2:
The patent changes the sampling parameter from uniform complete sampling to random undersampling with a defined probability p. This parameter change allows the system to acquire fewer data points while maintaining image quality through the statistical properties of random sampling, effectively reducing acquisition time without sacrificing measurement precision
2Loss of time
If k-space is undersampled to accelerate acquisition, then acquisition time is reduced, but image quality deteriorates with artifacts
Solution Approach 1:
By performing partial sampling of k-space points with random selection and probability p, the method achieves sufficient data for image reconstruction without requiring complete sampling. This partial action reduces acquisition time while the random sampling strategy preserves image quality by avoiding systematic aliasing artifacts
Solution Approach 2:
Instead of systematically sampling k-space in a regular grid pattern, the patent inverts the approach by randomly selecting which points to sample. This inversion from systematic to random sampling changes the artifact characteristics from regular aliasing patterns to noise-like artifacts that can be more effectively managed during reconstruction
3Productivity
If high undersampling rates are applied, then acquisition speed increases, but complex interpolation processes are required
Solution Approach 1:
The patent performs partial sampling at high undersampling rates while maintaining relatively simple reconstruction procedures. By randomly sampling k-space points and using straightforward reconstruction algorithms, the method achieves high acquisition speed without requiring complex interpolation or iterative reconstruction processes
Data Source
AI summary
In a method for the acquisition of magnetic resonance (MR) data relating to a pre-determined two-dimensional volume segment of an examination object with an MR apparatus, a randomized determination of points to be sampled in the raw data space is made, such that the raw data space is undersampled when only the determined points to be sampled are then sampled. MR data relating to the specified points to be sampled are acquired by operation of the MR apparatus. Alternatively, a determination of points to be sampled in the raw data space is made using radial or spiral trajectories in k-space that begin in the center k-space. Each specified point to be sampled is then moved to an FFT grid point, and MR data relating to the determined points to be sampled is implemented by operation of the MR apparatus.