Learnable slice-to-volume reconstruction for motion compensation in fetal magnetic resonance imaging
File(s)
Accepted version
Author(s)
Jehn, Constantin
Muller, Johanna P
Kainz, Bernhard
Type
Conference Paper
Abstract
Reconstructing motion-free 3D magnetic resonance imaging (MRI) volumes of fetal organs comes with the challenge of motion artefacts due to fetal motion and maternal respiration. Current methods rely on iterative procedures of outlier removal, super-resolution (SR) and slice-to-volume registration (SVR). Long runtimes and missing volume preservation over multiple iterations are still challenges for widespread clinical implementation. We envision an end-to-end learnable reconstruction framework that enables faster inference times and that can be steered by downstream tasks like segmentation. Therefore, we propose a new hybrid architecture for fetal brain reconstruction, consisting of a fully differentiable pre-registration module and a CycleGAN model for 3D image-toimage translation that is pretrained on our custom-generated dataset of 209 pairs of low-resolution (LoRes) and high-resolution (HiRes) fetal brain volumes. Our results are evaluated quantitatively with respect to five different similarity metrics. We incorporate the learned perceptual image patch similarity (LPIPS) metric and apply it to quantify volumetric image similarity for the first time in literature. Furthermore, we evaluate the model outputs qualitatively and conduct an expert survey to compare our method’s reconstruction quality to an established approach.
Editor(s)
Deserno, TM
Handels, H
Maier, A
Maier-Hein, K
Palm, C
Tolxdorff, T
Date Issued
2023-06-02
Date Acceptance
2023-07-01
Citation
Bildverarbeitung für die Medizin 2023, 2023, pp.25-31
ISBN
978-3-658-41656-0
ISSN
1431-472X
Publisher
Springer Vieweg Verlag
Start Page
25
End Page
31
Journal / Book Title
Bildverarbeitung für die Medizin 2023
Copyright Statement
© 2023 The Author(s), under exclusive license to Springer Nature Switzerland AG. This version of the article has been accepted for publication, after peer review (when applicable) and is subject to Springer Nature’s AM terms of use, but is not the Version of Record and does not reflect post-acceptance improvements, or any corrections. The Version of Record is available online at: http://dx.doi.org/10.1007/978-3-658-41657-7_10
Source
2023 German Conference on Medical Image Computing-BVM-Annual
Subjects
Computer Science
Computer Science, Artificial Intelligence
Engineering
Engineering, Biomedical
Life Sciences & Biomedicine
Radiology, Nuclear Medicine & Medical Imaging
Science & Technology
Technology
Publication Status
Published
Start Date
2023-07-02
Finish Date
2023-07-04
Coverage Spatial
Braunschweig, Germany
Date Publish Online
2023-06-02