EUVIMEDObjevte EUVIMED
Crossref

Low‐dose CT denoising using a diffusion prior via score distillation sampling

Hojin Jung, Jooho Lee, Jongduk Baek

Medical Physics · 2026

DOI: 10.1002/mp.70707 ↗

At a glance

Journal article in Medical Physics, 2026. Written by Hojin Jung and 2 co-authors. Institution: Department of Artificial Intelligence, Yonsei University, Seoul, Republic of Korea. Major MEDLINE topics: Tomography, X-Ray Computed, Radiation Dosage, Signal-To-Noise Ratio, Image Processing, Computer-Assisted. 50 references. Funded by: National Research Foundation of Korea, Institute for Informationand communications Technology.

AbstraktSubjectsBibliografické údajeCiteRelated studies

Abstract (original language)

Background

Low‐dose computed tomography (LDCT) reduces radiation exposure but amplifies noise, which can degrade diagnostic reliability. Recent studies have applied diffusion‐based models to LDCT denoising and shown promising performance. However, many of these approaches require paired LDCT and normal‐dose CT (NDCT) data or dose‐specific retraining, and their reconstruction procedures are often strictly coupled to a predefined diffusion trajectory.

Purpose

This study proposes an optimization‐based LDCT denoising framework that uses a pre‐trained NDCT diffusion model as an image prior and can be applied across various dose levels without dose‐specific retraining.

Methods

We formulate LDCT denoising as an optimization problem guided by a pre‐trained score‐based generative model (SGM) trained only on NDCT images. The Poisson data fidelity term enforces consistency with the measured projection data, whereas the score distillation sampling (SDS) term acts as a regularizer that guides the reconstruction toward the NDCT distribution learned by the pre‐trained SGM. The denoised image is obtained by iteratively optimizing a projection‐domain residual variable using the combined Poisson data fidelity gradient and SDS‐based regularization gradient.

Results

The proposed method effectively suppressed noise while preserving fine anatomical structures across multiple dose levels using a single pre‐trained diffusion model. Quantitative evaluations with respect to full‐dose reference images and visual comparisons demonstrated competitive performance compared with supervised and self‐supervised LDCT denoising baselines.

Conclusions

The results show that diffusion‐based image priors can be leveraged across various low‐dose levels within the proposed optimization‐based denoising framework. Our work suggests that SDS‐guided diffusion priors are a promising direction for LDCT denoising in data‐limited settings.

Subjects

MeSH terms (MEDLINE)

Tomography, X-Ray ComputedRadiation DosageSignal-To-Noise RatioImage Processing, Computer-AssistedDiffusionHumans

Author keywords

LDCTdiffusion priorimage denoisingscore distillation sampling

Authors and institutions

  • Department of Artificial Intelligence, Yonsei University, Seoul, Republic of Korea.
  • Bareunex Imaging, Inc., Seoul, Republic of Korea.

Funding

  • National Research Foundation of Korea
  • Institute for Informationand communications Technology

Cite

APA

Jung, H., Lee, J., & Baek, J. (2026). Low‐dose CT denoising using a diffusion prior via score distillation sampling. Medical Physics. https://doi.org/10.1002/mp.70707

Vancouver

Jung H, Lee J, Baek J. Low‐dose CT denoising using a diffusion prior via score distillation sampling. Medical Physics. 2026. doi:10.1002/mp.70707

BibTeX
@article{Jung2026,
  author = {Jung, Hojin and Lee, Jooho and Baek, Jongduk},
  title = {Low‐dose CT denoising using a diffusion prior via score distillation sampling},
  journal = {Medical Physics},
  year = {2026},
  doi = {10.1002/mp.70707}
}
Start for free →