Orientation-Aware Diffusion Super-Resolution for 3T-Like Fetal MRI from Routine 1.5T Scans

Xinliu Zhong, Ruiying Liu, Guohao Lin, Chuan Huang, Adam Ezra Goldman-Yassen, Amy Robben Mehollin-Ray, Yun Wang
Proceedings of The 9th International Conference on Medical Imaging with Deep Learning, PMLR 315:3827-3845, 2026.

Abstract

Fetal MRI plays a central role in assessing early brain development. While 3T scanners offer higher SNR and improved cortical detail, their increased sensitivity to motion, susceptibility artifacts, and $B_1$ inhomogeneity limits wide adoption for routine fetal imaging. Consequently, most clinical examinations are performed at 1.5T, where greater motion tolerance comes at the cost of lower SNR, reduced gray-white matter contrast, and partial-volume blurring - factors that undermine downstream morphometric analysis. Bridging this quality gap without sacrificing motion robustness of 1.5T would enable 3T-like morphometric reliability in routine clinical acquisitions. We propose an orientation-aware diffusion super-resolution framework that synthesizes 3T-like fetal brain contrast from routine 1.5T scans. The model combines a Swin-UNet backbone with gated FiLM-based orientation embeddings and a residual error-shifting diffusion mechanism. Training leverages the FaBiAN phantom to generate controllable high-/low-resolution pairs with monotonic intensity remapping, geometric perturbations, and simulated signal voids, thereby ensuring generalization to clinical data. Our model produces markedly sharper gyri and mitigates partial-volume effects in both synthesized and clinical data. When evaluated using Fetal-SynthSeg following NeSVoR reconstruction, the framework consistently improves tissue segmentation accuracy over state-of-the-art restoration baselines, yielding more reliable morphometric estimates for fetal brain analysis.

Cite this Paper

BibTeX
@InProceedings{pmlr-v315-zhong26a, title = {Orientation-Aware Diffusion Super-Resolution for 3T-Like Fetal MRI from Routine 1.5T Scans}, author = {Zhong, Xinliu and Liu, Ruiying and Lin, Guohao and Huang, Chuan and Goldman-Yassen, Adam Ezra and Mehollin-Ray, Amy Robben and Wang, Yun}, booktitle = {Proceedings of The 9th International Conference on Medical Imaging with Deep Learning}, pages = {3827--3845}, year = {2026}, editor = {Huo, Yuankai and Gao, Mingchen and Kuo, Chang-Fu and Jin, Yueming and Deng, Ruining}, volume = {315}, series = {Proceedings of Machine Learning Research}, month = {08--10 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v315/main/assets/zhong26a/zhong26a.pdf}, url = {https://proceedings.mlr.press/v315/zhong26a.html}, abstract = {Fetal MRI plays a central role in assessing early brain development. While 3T scanners offer higher SNR and improved cortical detail, their increased sensitivity to motion, susceptibility artifacts, and $B_1$ inhomogeneity limits wide adoption for routine fetal imaging. Consequently, most clinical examinations are performed at 1.5T, where greater motion tolerance comes at the cost of lower SNR, reduced gray-white matter contrast, and partial-volume blurring - factors that undermine downstream morphometric analysis. Bridging this quality gap without sacrificing motion robustness of 1.5T would enable 3T-like morphometric reliability in routine clinical acquisitions. We propose an orientation-aware diffusion super-resolution framework that synthesizes 3T-like fetal brain contrast from routine 1.5T scans. The model combines a Swin-UNet backbone with gated FiLM-based orientation embeddings and a residual error-shifting diffusion mechanism. Training leverages the FaBiAN phantom to generate controllable high-/low-resolution pairs with monotonic intensity remapping, geometric perturbations, and simulated signal voids, thereby ensuring generalization to clinical data. Our model produces markedly sharper gyri and mitigates partial-volume effects in both synthesized and clinical data. When evaluated using Fetal-SynthSeg following NeSVoR reconstruction, the framework consistently improves tissue segmentation accuracy over state-of-the-art restoration baselines, yielding more reliable morphometric estimates for fetal brain analysis.} }
Endnote
%0 Conference Paper %T Orientation-Aware Diffusion Super-Resolution for 3T-Like Fetal MRI from Routine 1.5T Scans %A Xinliu Zhong %A Ruiying Liu %A Guohao Lin %A Chuan Huang %A Adam Ezra Goldman-Yassen %A Amy Robben Mehollin-Ray %A Yun Wang %B Proceedings of The 9th International Conference on Medical Imaging with Deep Learning %C Proceedings of Machine Learning Research %D 2026 %E Yuankai Huo %E Mingchen Gao %E Chang-Fu Kuo %E Yueming Jin %E Ruining Deng %F pmlr-v315-zhong26a %I PMLR %P 3827--3845 %U https://proceedings.mlr.press/v315/zhong26a.html %V 315 %X Fetal MRI plays a central role in assessing early brain development. While 3T scanners offer higher SNR and improved cortical detail, their increased sensitivity to motion, susceptibility artifacts, and $B_1$ inhomogeneity limits wide adoption for routine fetal imaging. Consequently, most clinical examinations are performed at 1.5T, where greater motion tolerance comes at the cost of lower SNR, reduced gray-white matter contrast, and partial-volume blurring - factors that undermine downstream morphometric analysis. Bridging this quality gap without sacrificing motion robustness of 1.5T would enable 3T-like morphometric reliability in routine clinical acquisitions. We propose an orientation-aware diffusion super-resolution framework that synthesizes 3T-like fetal brain contrast from routine 1.5T scans. The model combines a Swin-UNet backbone with gated FiLM-based orientation embeddings and a residual error-shifting diffusion mechanism. Training leverages the FaBiAN phantom to generate controllable high-/low-resolution pairs with monotonic intensity remapping, geometric perturbations, and simulated signal voids, thereby ensuring generalization to clinical data. Our model produces markedly sharper gyri and mitigates partial-volume effects in both synthesized and clinical data. When evaluated using Fetal-SynthSeg following NeSVoR reconstruction, the framework consistently improves tissue segmentation accuracy over state-of-the-art restoration baselines, yielding more reliable morphometric estimates for fetal brain analysis.
APA
Zhong, X., Liu, R., Lin, G., Huang, C., Goldman-Yassen, A.E., Mehollin-Ray, A.R. & Wang, Y.. (2026). Orientation-Aware Diffusion Super-Resolution for 3T-Like Fetal MRI from Routine 1.5T Scans. Proceedings of The 9th International Conference on Medical Imaging with Deep Learning, in Proceedings of Machine Learning Research 315:3827-3845 Available from https://proceedings.mlr.press/v315/zhong26a.html.

Related Material