MRI Reconstruction with Selective SEMAC Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Magnetic resonance imaging (MRI) reconstruction is compromised by noise objects, such as metal implants, which cause field distortions, leading to artifacts and reduced image quality due to the need for extensive additional phase-encoding steps in existing correction methods like SEMAC, resulting in low spatial resolution and image degradation.
Innovation Solution
A method that differentiates between correction areas with distortions and standard reconstruction areas without distortions, applying SEMAC correction only where necessary, using a distortion criterion to selectively assign slice data for improved image quality by distinguishing signal and noise components and optimizing reconstruction techniques for each area.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If SEMAC correction is applied to correct metal-induced field distortions, then artifact reduction is improved, but measurement time increases significantly due to multiple additional phase-encoding steps
Solution Approach 1:
The patent applies SEMAC correction selectively only to image regions affected by metal artifacts (distorted regions), while leaving unaffected regions to be reconstructed using standard methods. This local differentiation allows artifact correction where needed without incurring the full time penalty of global SEMAC processing.
Solution Approach 2:
The image is segmented into distorted regions (requiring SEMAC correction) and non-distorted regions (amenable to standard reconstruction). This segmentation enables differential processing strategies that optimize both artifact correction and measurement time efficiency.
2Measurement precision
If additional phase-encoding steps are performed to resolve distorted excitation profiles, then spatial assignment accuracy is improved, but spatial resolution deteriorates due to low-resolution Fourier transforms
Solution Approach 1:
High-resolution spatial assignment is applied only to distorted regions where it is necessary for accurate artifact correction, while non-distorted regions maintain their original high spatial resolution. This localized application eliminates the trade-off in unaffected areas.
Solution Approach 2:
Instead of applying partial phase-encoding steps globally (which would reduce resolution), the patent applies full phase-encoding only locally where distortion exists, maintaining high resolution elsewhere while still achieving accurate spatial assignment where needed.
3Reliability
If slice data from adjacent partition slices are used for SEMAC reconstruction, then distortion correction is improved, but signal-to-noise ratio deteriorates due to inclusion of noise from outside the target slice
Solution Approach 1:
The patent selectively uses slice data from adjacent partition slices only in distorted regions where distortion correction is beneficial, while excluding such data from non-distorted regions to preserve signal-to-noise ratio. This local differentiation resolves the contradiction between correction effectiveness and noise introduction.
Solution Approach 2:
The patent identifies and utilizes the distorted excitation profile itself as a marker to guide where SEMAC correction should be applied. The harmful distortion effect becomes the basis for intelligent, localized correction decisions that avoid unnecessary noise introduction in clean regions.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach enhances image quality by reducing artifacts and maintaining high signal-to-noise ratio, avoiding the limitations of low-resolution Fourier transforms and signal overshooting, while minimizing the impact of SEMAC correction on undistorted regions.
Implementation Method 1
the object is normally introduced into a constant magnetic field (B0 field) of the magnetic resonance device, specifically into a homogeneity volume thereof in which only very slight deviations from the nominal value of the constant magnetic field are permitted. This causes nuclear spins in the target region to be oriented in the direction of the constant magnetic field.
Implementation Method 2
the spins are excited (deflected from alignment with the constant magnetic field) by radio-frequency pulses (excitation pulses) generated by a radio-frequency coil arrangement magnetic resonance signals produced by the decay of this excitation are detected
Implementation Method 3
In order to enable a spatial encoding of the magnetic resonance data, rapidly switched magnetic gradient fields are superimposed on the constant magnetic field, in particular a slice selection gradient that restricts the excitation to one slice to be measured, a phase encoding gradient, and/or a readout gradient
Implementation Method 4
The raw data present in k-space are converted by a Fourier transformation into the image domain in order to reconstruct a magnetic resonance image data set therefrom.
Data Source
AI summary
In a method for reconstruction of a three-dimensional image data set from magnetic resonance slice data of a target region acquired in target slices while a noise object distorting the magnetic field is present in the target region, for each target slice to be acquired, in addition to a central partition slice corresponding to the respective target slice, location, multiple partition slices adjacent to the central partition slice are acquired in a supplementary encoding direction perpendicular to the slice plane in multiple phase-encoding steps. A correction area and a standard reconstruction area of the target region are determined on the basis of a distortion criterion, obtained by evaluating the slice data that describes the distortion along the supplementary encoding direction. In the standard reconstruction area, only slice data are used, and in the correction area, slice data of partition slices outside the target slice are assigned to target slices in order to correct the distortion.