EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Optimized reconstruction of undersampled Dixon sequences using new memory‐efficient unrolled deep neural networks: HalfVarNet and HalfDIRCN

Sandra Martin iD, Amira Trabelsi, Maxime Guye, Marc Dubois, Redha Abdeddaim, David Bendahan, Rémi André

Magnetic Resonance in Medicine · 2026

DOI: 10.1002/mrm.70070 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Magnetic Resonance in Medicine, 2026. Verfasst von Sandra Martin und 6 weiteren Autor:innen. Beteiligte Einrichtung: Multiwave Technologies SAS, Marseille, France. Hauptthemen laut MEDLINE: Magnetic Resonance Imaging, Neural Networks, Computer, Image Processing, Computer-Assisted, Neuromuscular Diseases. Die Arbeit nennt 47 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: AMIDEX, ANRT, Association Nationale de la Recherche et de la Technologie.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Purpose

Fat fraction (FF) quantification in individual muscles using quantitative MRI is of major importance for monitoring disease progression and assessing disease severity in neuromuscular diseases. Undersampling of MRI acquisitions is commonly used to reduce scanning time. The present paper introduces novel unrolled neural networks for the reconstruction of undersampled MRI acquisitions. These networks are designed with the aim of maintaining accurate FF quantification while reducing reconstruction time and memory usage.

Methods

The proposed approach relies on a combination of a simplified architecture (Half U‐Net) with unrolled networks that achieved high performance in the well‐known FastMRI challenge (variational network [ VarNet] and densely interconnected residual cascading network [ DIRCN] ). The algorithms were trained and evaluated using 3D MRI Dixon acquisitions of the thigh from controls and patients with neuromuscular diseases. The study was performed by applying a retrospective undersampling with acceleration factors of 4 and 8. Reconstructed images were used to computed FF maps.

Results

Results disclose that the novel unrolled neural networks were able to maintain reconstruction, biomarker assessment, and segmentation quality while reducing memory usage by 24% to 16% and reducing reconstruction time from 21% to 17%. Using an acceleration factor of 8, the proposed algorithms, HalfVarNet and HalfDIRCN , achieved structural similarity index ( SSIM) scores of 93.76 ± 0.38 and 94.95 ± 0.32, mean squared error ( MSE) values of 12.76 ± 1.08 × 10 −2 and 10.25 ± 0.87 × 10 −2 , and a relative FF quadratic error of 0.23 ± 0.02% and 0.17 ± 0.02%, respectively.

Conclusion

The proposed method enables time and memory‐efficient reconstruction of undersampled 3D MRI data, supporting its potential for clinical application.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Magnetic Resonance ImagingNeural Networks, ComputerImage Processing, Computer-AssistedNeuromuscular DiseasesMuscle, SkeletalHumansAlgorithmsImaging, Three-DimensionalDeep LearningThighMaleAdultAdipose TissueFemaleImage Interpretation, Computer-Assisted

Stichwörter der Autor:innen

MRI accelerationVariational networkbiomarkermemory‐efficient learningneuromuscular disordersunrolled reconstruction

Autor:innen und Institutionen

  • Multiwave Technologies SAS, Marseille, France.
  • Aix Marseille Univ, CNRS, CRMBM, Marseille, France.
  • Aix Marseille Univ, CNRS, Centrale Med, Institut Fresnel, Marseille, France.

ORCID-Profile: Sandra Martin

Förderung

  • AMIDEX
  • ANRT
  • Association Nationale de la Recherche et de la Technologie

Zitieren

APA

Martin, S., Trabelsi, A., Guye, M., Dubois, M., Abdeddaim, R., Bendahan, D., & André, R. (2026). Optimized reconstruction of undersampled Dixon sequences using new memory‐efficient unrolled deep neural networks: HalfVarNet and HalfDIRCN. Magnetic Resonance in Medicine. https://doi.org/10.1002/mrm.70070

Vancouver

Martin S, Trabelsi A, Guye M, Dubois M, Abdeddaim R, Bendahan D, et al. Optimized reconstruction of undersampled Dixon sequences using new memory‐efficient unrolled deep neural networks: HalfVarNet and HalfDIRCN. Magnetic Resonance in Medicine. 2026. doi:10.1002/mrm.70070

BibTeX
@article{Martin2026,
  author = {Martin, Sandra and Trabelsi, Amira and Guye, Maxime and Dubois, Marc and Abdeddaim, Redha and Bendahan, David and André, Rémi},
  title = {Optimized reconstruction of undersampled Dixon sequences using new memory‐efficient unrolled deep neural networks: HalfVarNet and HalfDIRCN},
  journal = {Magnetic Resonance in Medicine},
  year = {2026},
  doi = {10.1002/mrm.70070}
}
Kostenlos starten →