SRNR: Deep learning-based MRI super-resolution using noisy high-resolution reference data
Imaging Neuroscience · 03.11.2026 · 4
DOI: 10.1162/imag.a.1433 ↗Kurz gefasst
Fachartikel in der Zeitschrift Imaging Neuroscience, online erschienen am 03.11.2026. Verfasst von Muye Zhang und 17 weiteren Autor:innen. Beteiligte Einrichtung: School of Biomedical Engineering, Tsinghua University, Beijing, China. Hauptthemen laut MEDLINE: Deep Learning, Magnetic Resonance Imaging, Image Processing, Computer-Assisted, Brain. Die Arbeit nennt 61 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Natural Science Foundation of China, Tsinghua University.
Abstract (Originalsprache)
Deep learning-based super-resolution (SR) techniques offer a powerful tool to enhance the spatial resolution of magnetic resonance imaging (MRI) without increasing scan time. However, most supervised SR methods rely on high-resolution, high-signal-to-noise ratio (SNR) reference data for training, which are difficult to acquire in many clinical settings, such as in time-constrained pediatric imaging. Inspired by the Noise2Noise principle, we propose SRNR, a new training paradigm that enables supervised SR using noisy high-resolution images as references. SRNR was systematically evaluated on both pediatric and adult simulated data across different noise levels of the reference data. SRNR generated super-resolved image volumes with high similarity to the high-resolution references (PSNR: 32.5–35.2 dB) and improved downstream image analysis performance, including tissue segmentation (white matter Dice coefficient: 0.951–0.960) and cortical surface reconstruction (cortical thickness absolute difference: 0.200–0.255 mm). Furthermore, SRNR was shown to be effective on clinically acquired pediatric brain MRI data, achieving improvements consistent with those observed in the simulated experiments. These results demonstrate that SRNR substantially reduces reliance on clean, high-SNR training data and enables effective SR under clinically realistic acquisition constraints, thereby improving the feasibility of deploying deep learning-based SR in pediatric MRI and related applications.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- School of Biomedical Engineering, Tsinghua University, Beijing, China.
- Department of Radiology, West China Second University Hospital, Sichuan University, Chengdu, China.
- MR Research Collaborations, Siemens Healthineers Ltd., Chengdu, China.
- Siemens Healthineers AG, Erlangen, Germany.
- Department of Radiology and Biomedical Imaging, University of California San Francisco, San Francisco, CA, United States.
- Athinoula A. Martinos Center for Biomedical Imaging, Department of Radiology, Massachusetts General Hospital, Charlestown, MA, United States.
ORCID-Profile: Hongjia Yang, Jiaxin Xiao, Yuhang He, Ting Yin, Jialan Zheng, Mingxuan Liu, Wei Liu, Congyu Liao, Berkin Bilgic, Jonathan R Polimeni, Susie Y Huang, Ziyu Li, Haibo Qu, Qiyuan Tian
Förderung
- National Natural Science Foundation of China
- Tsinghua University
Zitieren
APA
Zhang, M., Yang, H., Li, Z., Xiao, J., Liu, W., He, Y., Yin, T., Zheng, J., Liu, M., Liu, W., Liao, C., Bilgic, B., Polimeni, J. R., Huang, S. Y., Liao, Y., Li, Z., Qu, H., & Tian, Q. (2026). SRNR: Deep learning-based MRI super-resolution using noisy high-resolution reference data. Imaging Neuroscience, 4. https://doi.org/10.1162/imag.a.1433
Vancouver
Zhang M, Yang H, Li Z, Xiao J, Liu W, He Y, et al. SRNR: Deep learning-based MRI super-resolution using noisy high-resolution reference data. Imaging Neuroscience. 2026;4. doi:10.1162/imag.a.1433
BibTeX
@article{Zhang2026,
author = {Zhang, Muye and Yang, Hongjia and Li, Zihan and Xiao, Jiaxin and Liu, Wanqing and He, Yuhang and Yin, Ting and Zheng, Jialan and Liu, Mingxuan and Liu, Wei and Liao, Congyu and Bilgic, Berkin and Polimeni, Jonathan R and Huang, Susie Y and Liao, Yi and Li, Ziyu and Qu, Haibo and Tian, Qiyuan},
title = {SRNR: Deep learning-based MRI super-resolution using noisy high-resolution reference data},
journal = {Imaging Neuroscience},
year = {2026},
volume = {4},
doi = {10.1162/imag.a.1433}
}
EUVIMED