Self‐Supervised Noise Adaptive MRI Denoising via Repetition to Repetition (Rep2Rep) Learning
Magnetic Resonance in Medicine · 2026-03 · 95 (3) · 1619-1633
DOI: 10.1002/mrm.70155 ↗Kurz gefasst
Fachartikel in der Zeitschrift Magnetic Resonance in Medicine, online erschienen am 09.11.2025. Verfasst von Nikola Janjušević und 7 weiteren Autor:innen. Beteiligte Einrichtung: Bernard and Irene Schwartz Center for Biomedical Imaging, Department of Radiology, New York University Grossman School of Medicine, New York, New Yor…. Hauptthemen laut MEDLINE: Magnetic Resonance Imaging, Image Processing, Computer-Assisted, Supervised Machine Learning. Die Arbeit wurde bisher 11-mal zitiert (Crossref) und nennt 33 Literaturangaben. Der Volltext ist frei in PubMed Central verfügbar. Gefördert von: NIBIB NIH HHS, NIDDK NIH HHS, National Institute of Biomedical Imaging and Bioengineering.
Abstract (Originalsprache)
Purpose
This work proposes a novel self‐supervised noise‐adaptive image denoising framework, called Repetition to Repetition (Rep2Rep) learning, for low‐field ( 1T) MRI applications.
Methods
Rep2Rep learning extends the Noise2Noise framework by training a neural network on two repeated MRI acquisitions, using one repetition as input and another as target, without requiring ground‐truth data. It incorporates noise‐adaptive training, enabling denoising generalization across varying noise‐levels and flexible inference with any number of repetitions. Performance was evaluated on both synthetic noisy Brain MRI and 0.55T Prostate MRI data, and compared against supervised learning and Monte Carlo Stein's Unbiased Risk Estimator (MC‐SURE).
Results
Rep2Rep learning outperforms MC‐SURE on both synthetic and 0.55T MRI datasets. On synthetic Brain data, it achieved denoising quality comparable to supervised learning and surpassed MC‐SURE, particularly in preserving structural details and reducing residual noise. On the 0.55T Prostate MRI data a reader study showed that Rep2Rep‐denoised 2‐average images outperformed 8‐average noisy images. Rep2Rep demonstrated robustness to noise‐level discrepancies between training and inference, supporting its practical implementation.
Conclusion
Rep2Rep learning offers an effective self‐supervised denoising for low‐field MRI by leveraging routinely acquired multi‐repetition data. Its noise‐adaptivity enables generalization to different SNR regimes without clean reference images. This makes Rep2Rep learning a promising tool for improving image quality and scan efficiency in low‐field MRI.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Bernard and Irene Schwartz Center for Biomedical Imaging, Department of Radiology, New York University Grossman School of Medicine, New York, New York, USA.
- Center for Advanced Imaging Innovation and Research (CAI2R), Department of Radiology, New York University Grossman School of Medicine, New York, New York, USA.
- Video Lab, Department of Electrical and Computer Engineering, New York University Tandon School of Engineering, Brooklyn, New York, USA.
ORCID-Profile: Nikola Janjušević, Jingjia Chen, Yao Wang, Li Feng
Förderung
- NIBIB NIH HHS
- NIDDK NIH HHS
- National Institute of Biomedical Imaging and Bioengineering
Zitieren
APA
Janjušević, N., Chen, J., Ginocchio, L., Bruno, M., Huang, Y., Wang, Y., Chandarana, H., & Feng, L. (2026). Self‐Supervised Noise Adaptive MRI Denoising via Repetition to Repetition (Rep2Rep) Learning. Magnetic Resonance in Medicine, 95(3), 1619-1633. https://doi.org/10.1002/mrm.70155
Vancouver
Janjušević N, Chen J, Ginocchio L, Bruno M, Huang Y, Wang Y, et al. Self‐Supervised Noise Adaptive MRI Denoising via Repetition to Repetition (Rep2Rep) Learning. Magnetic Resonance in Medicine. 2026;95(3):1619-1633. doi:10.1002/mrm.70155
BibTeX
@article{Janjuevi2026,
author = {Janjušević, Nikola and Chen, Jingjia and Ginocchio, Luke and Bruno, Mary and Huang, Yuhui and Wang, Yao and Chandarana, Hersh and Feng, Li},
title = {Self‐Supervised Noise Adaptive MRI Denoising via Repetition to Repetition (Rep2Rep) Learning},
journal = {Magnetic Resonance in Medicine},
year = {2026},
volume = {95},
number = {3},
pages = {1619--1633},
doi = {10.1002/mrm.70155}
}
EUVIMED