Data Consistent Deep Rigid MRI Motion Correction

Nalini M Singh, Neel Dey, Malte Hoffmann, Bruce Fischl, Elfar Adalsteinsson, Robert Frost, Adrian V Dalca, Polina Golland
Medical Imaging with Deep Learning, PMLR 227:368-381, 2024.

Abstract

Motion artifacts are a pervasive problem in MRI, leading to misdiagnosis or mischaracterization in population-level imaging studies. Current retrospective rigid intra-slice motion correction techniques jointly optimize estimates of the image and the motion parameters. In this paper, we use a deep network to reduce the joint image-motion parameter search to a search over rigid motion parameters alone. Our network produces a reconstruction as a function of two inputs: corrupted k-space data and motion parameters. We train the network using simulated, motion-corrupted k-space data generated from known motion parameters. At test-time, we estimate unknown motion parameters by minimizing a data consistency loss between the motion parameters, the network-based image reconstruction given those parameters, and the acquired measurements. Intra-slice motion correction experiments on simulated and realistic 2D fast spin echo brain MRI achieve high reconstruction fidelity while retaining the benefits of explicit data consistency-based optimization.

Cite this Paper

BibTeX
@InProceedings{pmlr-v227-singh24a, title = {Data Consistent Deep Rigid MRI Motion Correction}, author = {Singh, Nalini M and Dey, Neel and Hoffmann, Malte and Fischl, Bruce and Adalsteinsson, Elfar and Frost, Robert and Dalca, Adrian V and Golland, Polina}, booktitle = {Medical Imaging with Deep Learning}, pages = {368--381}, year = {2024}, editor = {Oguz, Ipek and Noble, Jack and Li, Xiaoxiao and Styner, Martin and Baumgartner, Christian and Rusu, Mirabela and Heinmann, Tobias and Kontos, Despina and Landman, Bennett and Dawant, Benoit}, volume = {227}, series = {Proceedings of Machine Learning Research}, month = {10--12 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v227/singh24a/singh24a.pdf}, url = {https://proceedings.mlr.press/v227/singh24a.html}, abstract = {Motion artifacts are a pervasive problem in MRI, leading to misdiagnosis or mischaracterization in population-level imaging studies. Current retrospective rigid intra-slice motion correction techniques jointly optimize estimates of the image and the motion parameters. In this paper, we use a deep network to reduce the joint image-motion parameter search to a search over rigid motion parameters alone. Our network produces a reconstruction as a function of two inputs: corrupted k-space data and motion parameters. We train the network using simulated, motion-corrupted k-space data generated from known motion parameters. At test-time, we estimate unknown motion parameters by minimizing a data consistency loss between the motion parameters, the network-based image reconstruction given those parameters, and the acquired measurements. Intra-slice motion correction experiments on simulated and realistic 2D fast spin echo brain MRI achieve high reconstruction fidelity while retaining the benefits of explicit data consistency-based optimization.} }
Endnote
%0 Conference Paper %T Data Consistent Deep Rigid MRI Motion Correction %A Nalini M Singh %A Neel Dey %A Malte Hoffmann %A Bruce Fischl %A Elfar Adalsteinsson %A Robert Frost %A Adrian V Dalca %A Polina Golland %B Medical Imaging with Deep Learning %C Proceedings of Machine Learning Research %D 2024 %E Ipek Oguz %E Jack Noble %E Xiaoxiao Li %E Martin Styner %E Christian Baumgartner %E Mirabela Rusu %E Tobias Heinmann %E Despina Kontos %E Bennett Landman %E Benoit Dawant %F pmlr-v227-singh24a %I PMLR %P 368--381 %U https://proceedings.mlr.press/v227/singh24a.html %V 227 %X Motion artifacts are a pervasive problem in MRI, leading to misdiagnosis or mischaracterization in population-level imaging studies. Current retrospective rigid intra-slice motion correction techniques jointly optimize estimates of the image and the motion parameters. In this paper, we use a deep network to reduce the joint image-motion parameter search to a search over rigid motion parameters alone. Our network produces a reconstruction as a function of two inputs: corrupted k-space data and motion parameters. We train the network using simulated, motion-corrupted k-space data generated from known motion parameters. At test-time, we estimate unknown motion parameters by minimizing a data consistency loss between the motion parameters, the network-based image reconstruction given those parameters, and the acquired measurements. Intra-slice motion correction experiments on simulated and realistic 2D fast spin echo brain MRI achieve high reconstruction fidelity while retaining the benefits of explicit data consistency-based optimization.
APA
Singh, N.M., Dey, N., Hoffmann, M., Fischl, B., Adalsteinsson, E., Frost, R., Dalca, A.V. & Golland, P.. (2024). Data Consistent Deep Rigid MRI Motion Correction. Medical Imaging with Deep Learning, in Proceedings of Machine Learning Research 227:368-381 Available from https://proceedings.mlr.press/v227/singh24a.html.

Related Material