Reducing and correcting magnetic field gradient deviations

Patent metadata

Field	Value
Publication number	US-2023358834-A1
Application number	US-202318142154-A
Country	US
Kind code	A1
Filing date	May 2, 2023
Priority date	May 3, 2022
Publication date	Nov 9, 2023
Grant date	—

Abstract

In a method for reducing and/or correcting deviations from a target gradient of a magnetic field gradient created by an MR system input data is provided for a trained function trained by a machine-learning algorithm, wherein the input data comprises information about the target gradient of the MR system. The trained function further creates output data with the aid of the input data. The deviations from the target gradient of the magnetic field gradient created by the MR system are reduced and/or corrected with the aid of the output data created.

First claim

1 . A method for reducing, correcting, or reducing and correcting deviations from a target gradient of a magnetic field gradient created by a magnetic resonance (MR) system, the method comprising:
providing input data for a function trained by a machine-learning algorithm, wherein the input data comprises information about the target gradient of the MR system; creating output data by the trained function when input the input data; and reducing, correcting, or reducing and correcting the deviations from the target gradient of the magnetic field gradient created by the MR system using the output data.

2 . The method of claim 1 , wherein the MR system comprises an amplifier that is configured to amplify an amplifier input signal and output an amplifier output signal, and wherein the MR system further comprises a gradient coil that is configured to create the magnetic field gradient using at least the amplifier output signal; wherein the trained function determines a corrected amplifier input signal or a correction signal for correction of the amplifier input signal; and wherein the amplifier output signal is created as a function of the corrected amplifier input signal or of the correction signal.

3 . The method of claim 1 , wherein the trained function determines an estimation for a magnetic field gradient created by the MR system.

4 . The method of claim 3 , wherein a gradient characterization function is determined using the estimation for the magnetic field gradients created by the MR system.

5 . The method of claim 4 , further comprising:
creating a corrected amplifier input signal or a correction signal for correction of the amplifier input signal using the gradient characterization function; and applying the corrected amplifier input signal to the amplifier.

6 . The method of claim 5 , wherein the MR system creates raw data by the magnetic field gradient created; and wherein MR image data is created using the raw data and the gradient characterization function.

7 . The method of claim 1 , wherein the input data for the trained function further comprises at least one of:
an amplifier input signal of an amplifier of the MR system, wherein the amplifier amplifies the amplifier input signal and outputs it as the amplifier output signal to a gradient coil of the MR system, which creates a magnetic field gradient using the amplifier output signal, the measured amplifier output signal, at least one measured temperature of the amplifier, of the gradient coils, or of the amplifier and the of the gradient coils, a measured magnetic field gradient created by the MR system, a diagnostically relevant imaging region, or information regarding an object being examined by the MR system.

8 . A computer-implemented method for creation of a trained function, the method comprising:
providing training data, the training data comprising at least one item of information about a target gradient of a magnetic resonance (MR) system with associated magnetic field gradient created by the MR system; training the function using a machine-learning algorithm, based on the training data, wherein the function is trained to output output data that is able to be used for reducing, correcting, or reducing and correcting deviations from the target gradient of magnetic field gradients created by the MR system; and providing the trained function.

9 . The method of claim 8 , wherein the trained function is created such that an estimation for a magnetic field gradient created by the MR system is determined using the trained function, wherein a gradient characterization function is determined using the estimation for the magnetic field gradient created by the MR system, and
using the gradient characterization function a corrected amplifier input signal or a correction signal for correction of the amplifier input signal is created, or MR image data is created using raw data that is created by the MR system by the magnetic field gradient and is created using the gradient characterization function.

10 . The method of claim 8 , wherein the training data comprises information about at least one of the target gradient of the MR system and associated magnetic field gradients created by the MR system for a plurality of different temperatures of at least one component of the MR system, different frequencies, different pulse widths, different amplitudes, or different slew rates.

11 . The method of claim 8 , wherein the training data for a pair consists of the target gradient and an associated magnetic field gradient created by the MR system, wherein the training data further comprises at least one of
an amplifier input signal of an amplifier of the MR system, wherein the amplifier amplifies the amplifier input signal and outputs an amplifier output signal to gradient coils of the MR system that create a magnetic field gradient using the amplifier output signal, the amplifier output signal, at least one measured temperature of the amplifier, of the gradient coils, or of the amplifier and the gradient coils, a measured magnetic field gradient created by the MR system, an image region, or information related to an object being examined by the MR system.

12 . An apparatus for reducing, correcting, or reducing and correcting deviations from a target gradient of a magnetic field gradient created by a magnetic resonance (MR) system, the apparatus comprising:
a computing facility configured to provide input data to a trained function trained by a machine-learning algorithm, wherein the input data comprises information about the target gradient of the MR system, and to create output data by the trained function when input the input data; and a reduction/correction facility configured using the output data of the trained function to reduce, correct, or reduce and correct deviations from the target gradient of the magnetic field gradient created by the MR system.

13 . The apparatus of claim 12 , wherein the MR system comprises:
an amplifier configured to amplify an amplifier input signal and to output an amplifier output signal; and a gradient coil configured to create a magnetic field gradient using the amplifier output signal.

Assignees

• Siemens Healthcare Gmbh

Inventors

• Stich Manuel
• Puls Philipp

Classifications

• G01R33/389 Primary

  Field stabilisation {, e.g. by field measurements and control means or indirectly by current stabilisation} · CPC title

• G01R33/56572 Primary

  caused by a distortion of a gradient magnetic field, e.g. non-linearity of a gradient magnetic field (G01R33/56509, G01R33/56518, G01R33/56536 take precedence) · CPC title

• G01R33/5608

  Data processing and visualization specially adapted for MR, e.g. for feature analysis and pattern recognition on the basis of measured MR data, segmentation of measured MR data, edge contour detection on the basis of measured MR data, for enhancing measured MR data in terms of signal-to-noise ratio by means of noise filtering or apodization, for enhancing measured MR data in terms of resolution by means for deblurring, windowing, zero filling, or generation of gray-scaled images, colour-coded images or images displaying vectors instead of pixels (image data processing or generation, in general G06T) · CPC title

• G06N3/02

  Neural networks · CPC title

• G06N20/00

  Machine learning · CPC title