System and method for reconstruction using a high-resolution phase in magnetic resonance images
A method for producing an image of a subject with a magnetic resonance imaging (MRI) comprises acquiring a first set of partial k-space data from the subject and generating a phase corrected image based on a phase correction factor and the first set of the partial k-space data. The method further includes transforming the phase corrected image into a second set of partial k-space data and reconstructing the image of the subject from the second set of the partial k-space data and a weighting function.
1. A method for producing an image of a subject with a magnetic resonance imaging (MRI) system, the method comprising:
acquiring, using the MRI system, a first set of partial k-space data from the subject, wherein the first set of the partial k-space data is asymmetric and truncated in at least one k-space dimension;
zero-filling the first set of the partial k-space data with zeros at locations corresponding to skipped k-space locations to derive full k-space data;
reconstructing a first image based on the zero-filled k-space data;
determining a phase correction factor from the first image;
generating a phase corrected image based on the phase correction factor and the first set of the partial k-space data;
transforming the phase corrected image into a second set of partial k-space data; and
reconstructing, from the second set of the partial k-space data and a weighting function, the image of the subject.
2. The method of claim 1 , comprises using a neural network model with the first image to determine the phase correction factor.
3. The method of claim 2 , wherein the neural network model is trained with a pair of pristine images and corrupted images, wherein the corrupted images include corrupted phase information, the pristine images are the corrupted images with the corrupted phase information reduced.
4. The method of claim 2 , wherein the neural network model is trained with a loss function including a constraint based on prior knowledge of phase information.
5. The method of claim 1 , wherein generating the phase corrected image based on the phase correction factor and the first set of the partial k-space data comprises multiplying the first image with the phase correction factor.
6. The method of claim 1 , wherein the image of the subject is reconstructed by determining a real portion of a multiplication result of the weighting function and the second set of the partial k-space data.
7. The method of claim 1 , wherein the weighting function represents a spectral sampling density.
8. A magnetic resonance imaging (MRI) system, comprising:
a magnet configured to generate a polarizing magnetic field about at least a portion of a subject arranged in the MRI system;
a gradient coil assembly including a plurality of gradient coils configured to apply at least one gradient field to the polarizing magnetic field;
a radio frequency (RF) system configured to apply an RF field to the subject and to receive magnetic resonance signals from the subject; and
a processing system programmed to:
acquire a first set of partial k-space data from the subject, wherein the first set of partial k-space data is asymmetric and truncated in at least one k-space dimension;
zero-fill the first set of the partial k-space data with zeros at locations corresponding to skipped k-space locations to derive full k-space data;
reconstruct a first image based on the zero-filled k-space data;
determine a phase correction factor from the first image;
generate a phase corrected image based on the phase correction factor and the first set of the partial k-space data;
transform the phase corrected image into a second set of partial k-space data; and
reconstruct, from the second set of the partial k-space data and a weighting function, the image of the subject.
9. The MRI system of claim 8 , wherein the processing system is programmed to determine the phase correction factor from the first image using a neural network model.
10. The MRI system of claim 9 , wherein the processing system is programmed to train the neural network model with a pair of pristine images and corrupted images, wherein the corrupted images include corrupted phase information, the pristine images are the corrupted images with the corrupted phase information reduced.
11. A method of MR imaging, comprising:
acquiring a first set of k-space data for less than all of k-space from a subject, wherein the first set of the k-space data is asymmetric and truncated in at least one k-space dimension;
zero-filling the first set of the k-space data with zeros at locations corresponding to skipped k-space locations to derive full k-space data;
reconstructing a first image based on the zero-filled k-space data;
determining a phase correction factor from the first image;
generating the phase corrected image based on a phase correction factor and the first set of the k-space data;
transforming the phase corrected image into a second set of k-space data; and
reconstructing, from the second set of the k-space data and a weighting function, the image of the subject.
12. The method of claim 11 , comprising determining the phase correction factor from the first image using ene of a neural network model.
13. The method of claim 12 , wherein the neural network model is trained with a pair of pristine images and corrupted images, wherein the corrupted images include corrupted phase information, the pristine images are the corrupted images with the corrupted phase information reduced.