Radial MRI Oscillogram Filtering for RF Interference
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current MRI technologies are inadequate in mitigating intermittent and broadband noise artifacts in radial MRI data, particularly the 'zipper' artifact caused by radio frequency interference, which affects image quality.
Innovation Solution
The method involves transforming radial MRI data into the oscillogram domain, where noise artifacts appear as horizontal lines, allowing for filtering and suppression while preserving image data, using techniques such as integrating across the horizontal axis and applying a 1D projection to automatically recognize and exclude RF zipper artifacts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If current filtering technologies are applied to radial MRI data, then processing time is reduced, but noise artifacts like zipper artifacts cannot be effectively removed
Solution Approach 1:
The patent transforms the MRI data from the spatial domain to the oscillogram domain through Fourier transformation. In this transformed domain, noise artifacts such as zipper artifacts manifest as distinct horizontal lines at specific frequency positions, while genuine image data appears as vertical structures. This dimensional transformation enables selective filtering of artifacts while preserving image content.
Solution Approach 2:
The filtering process segments the oscillogram data into artifact components and image components based on their distinct patterns. Horizontal lines identified as artifacts are separated and removed, while vertical structures representing genuine image data are preserved. This segmentation approach allows selective removal of harmful artifacts without affecting image quality.
2Object-affected harmful factors
If aggressive filtering is applied to remove artifacts, then noise reduction is improved, but image data distortion increases
Solution Approach 1:
The filtering operation applies different treatments to different regions of the oscillogram based on local characteristics. Horizontal line regions (artifacts) are suppressed while vertical region structures (image data) are preserved. This localized quality approach ensures that filtering is applied only where needed without distorting genuine image information.
Solution Approach 2:
The system uses the structural characteristics in the oscillogram domain as feedback to guide the filtering process. By identifying horizontal lines as artifacts and vertical structures as image data, the system automatically adjusts filtering strength in different regions, preventing over-filtering that would distort image content while ensuring adequate artifact removal.
Data Source
AI summary
A system includes a data receiver, a sinogram generator, a processor, a filter module, and an output module. The data receiver is configured to receive radial ordered magnetic resonance data. The sinogram generator is configured to generate a first sinogram corresponding to a view angle as a function of a readout direction for the magnetic resonance data. The processor is configured to generate an oscillogram having an angular frequency axis. The oscillogram corresponds to a Fourier transform of the first sinogram. The filter module is configured to selectively filter a peak in a projection formed along a selected axis of the oscillogram, the peak being related to an interference signal such as an RF interference. The selected axis is orthogonal to the angular frequency axis. The output module is configured to form a second sinogram corresponding to a transform of the filtered projection.