EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Deep Guess acceleration for explainable image reconstruction in sparse-view CT

Elena Loli Piccolomini, Davide Evangelista, Elena Morotti

Computerized Medical Imaging and Graphics · 2025-07 · 123 · 102530

DOI: 10.1016/j.compmedimag.2025.102530 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Computerized Medical Imaging and Graphics, 2025-07. Verfasst von Elena Loli Piccolomini und 2 weiteren Autor:innen. Beteiligte Einrichtung: Department of Computer Science and Engineering, University of Bologna, Via Mura Anteo Zamboni 7, 40126 Bologna, Italy. Electronic address: elena.loli…. Hauptthemen laut MEDLINE: Tomography, X-Ray Computed, Deep Learning, Image Processing, Computer-Assisted, Radiographic Image Interpretation, Computer-Assisted. Die Arbeit wurde bisher 7-mal zitiert (Crossref) und nennt 67 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: European Union, Francesco Severi National Institute of Higher Mathematics National Group of Scientific Calculations.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Sparse-view Computed Tomography (CT) is an emerging protocol designed to reduce X-ray dose radiation in medical imaging. Reconstructions based on the traditional Filtered Back Projection algorithm suffer from severe artifacts due to sparse data. In contrast, Model-Based Iterative Reconstruction (MBIR) algorithms, though better at mitigating noise through regularization, are too computationally costly for clinical use. This paper introduces a novel technique, denoted as the Deep Guess acceleration scheme, using a trained neural network both to quicken the regularized MBIR and to enhance the reconstruction accuracy. We integrate state-of-the-art deep learning tools to initialize a clever starting guess for a proximal algorithm solving a non-convex model and thus computing a (mathematically) interpretable solution image in a few iterations. Experimental results on real and synthetic CT images demonstrate the Deep Guess effectiveness in (very) sparse tomographic protocols, where it overcomes its mere variational counterpart and many data-driven approaches at the state of the art. We also consider a ground truth-free implementation and test the robustness of the proposed framework to noise.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Tomography, X-Ray ComputedDeep LearningImage Processing, Computer-AssistedRadiographic Image Interpretation, Computer-AssistedHumansAlgorithmsNeural Networks, ComputerArtifactsPhantoms, ImagingRadiation Dosage

Stichwörter der Autor:innen

Deep neural networksInterpretable reconstructionModel-based iterative reconstructionNon-convex optimizationSparse view computed tomography

Autor:innen und Institutionen

  • Department of Computer Science and Engineering, University of Bologna, Via Mura Anteo Zamboni 7, 40126 Bologna, Italy. Electronic address: elena.loli@unibo.it.
  • Department of Computer Science and Engineering, University of Bologna, Via Mura Anteo Zamboni 7, 40126 Bologna, Italy.
  • Department of Political and Social Sciences, University of Bologna, Strada Maggiore 45, 40125 Bologna, Italy.

Förderung

  • European Union
  • Francesco Severi National Institute of Higher Mathematics National Group of Scientific Calculations

Zitieren

APA

Loli Piccolomini, E., Evangelista, D., & Morotti, E. (2025). Deep Guess acceleration for explainable image reconstruction in sparse-view CT. Computerized Medical Imaging and Graphics, 123, 102530. https://doi.org/10.1016/j.compmedimag.2025.102530

Vancouver

Loli Piccolomini E, Evangelista D, Morotti E. Deep Guess acceleration for explainable image reconstruction in sparse-view CT. Computerized Medical Imaging and Graphics. 2025;123:102530. doi:10.1016/j.compmedimag.2025.102530

BibTeX
@article{LoliPiccolomini2025,
  author = {Loli Piccolomini, Elena and Evangelista, Davide and Morotti, Elena},
  title = {Deep Guess acceleration for explainable image reconstruction in sparse-view CT},
  journal = {Computerized Medical Imaging and Graphics},
  year = {2025},
  volume = {123},
  pages = {102530},
  doi = {10.1016/j.compmedimag.2025.102530}
}
Kostenlos starten →