EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

FMDNet: Spatial‐frequency feature routing for low‐dose CT denoising

Yujie Yao, Yunsen Liang iD, Wen Xiao, Zibo Zhou, Yaoxue Xu, Yonghao Pan, Xiao Xia

Journal of Applied Clinical Medical Physics · 2026

DOI: 10.1002/acm2.70656 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Applied Clinical Medical Physics, 2026. Verfasst von Yujie Yao und 6 weiteren Autor:innen. Beteiligte Einrichtung: School of Information Engineering, Sichuan Agricultural University, Ya'an, Sichuan, China. Hauptthemen laut MEDLINE: Tomography, X-Ray Computed, Image Processing, Computer-Assisted, Algorithms, Deep Learning. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 53 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Provincial Undergraduate Training Program on Innovation and Entrepreneurship.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Background

Low‐dose computed tomography (LDCT) is widely used to reduce radiation exposure, but the reduced photon budget amplifies quantum noise and can introduce structured artifacts that obscure subtle boundaries and textures. Many deep learning denoisers process features in a single stream, which may encourage either over‐smoothing of weak anatomical edges or unstable texture synthesis.

Purpose

To develop an LDCT denoising network that explicitly routes coarse structural content and fine details through dedicated pathways, aiming to suppress noise while preserving anatomically meaningful high‐frequency information.

Methods

We propose FMDNet, a frequency‐aware encoder‐decoder equipped with an explicit coarse/detail routing block. A fixed low‐pass operator produces a coarse component, while the corresponding detail component is formed as an explicit residual (high = x—low). Each component is refined with depthwise operators and fused by a learned channel gate with channel attention. We evaluate FMDNet on the AAPM Mayo Clinic LDCT dataset and the LoDoPaB‐CT benchmark under a reproducible HU‐windowed protocol and supplement PSNR, SSIM, and RMSE with texture‐ and structure‐oriented analyses including HU line profiles and noise power spectrum (NPS), together with volume‐level paired tests and bootstrap confidence intervals on LoDoPaB‐CT.

Results

Across both benchmarks, FMDNet achieves competitive quantitative fidelity and shows favorable results on complementary structure‐/texture‐oriented analyses. On Mayo, it improves mean PSNR, SSIM, and RMSE relative to strong learned baselines. On LoDoPaB‐CT, volume‐level analysis across 28 validation volumes confirms consistent improvements over Uformer with paired tests and bootstrap confidence intervals. Additional HU‐profile and NPS analyses provide complementary evidence of improved texture preservation and local structural fidelity relative to comparative methods.

Conclusions

Explicitly separating coarse and detail residual components in feature space provides a practical inductive bias for LDCT denoising. When combined with gated fusion and multi‐scale supervision, this strategy improves quantitative fidelity and preserves fine structures without relying on adversarial texture synthesis; clinical diagnostic impact should be validated by reader studies.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Tomography, X-Ray ComputedImage Processing, Computer-AssistedAlgorithmsDeep LearningHumansRadiation DosageSignal-To-Noise RatioArtifacts

Stichwörter der Autor:innen

deep learningimage denoisinglow‐dose computed tomography (LDCT)

Autor:innen und Institutionen

  • School of Information Engineering, Sichuan Agricultural University, Ya'an, Sichuan, China.
  • College of Science, Sichuan Agricultural University, Ya'an, Sichuan, China.
  • Department of Radiology, Clinical Skills Center, Ya'an People's Hospital, Ya'an, Sichuan, China.

ORCID-Profile: Yunsen Liang

Förderung

  • Provincial Undergraduate Training Program on Innovation and Entrepreneurship

Zitieren

APA

Yao, Y., Liang, Y., Xiao, W., Zhou, Z., Xu, Y., Pan, Y., & Xia, X. (2026). FMDNet: Spatial‐frequency feature routing for low‐dose CT denoising. Journal of Applied Clinical Medical Physics. https://doi.org/10.1002/acm2.70656

Vancouver

Yao Y, Liang Y, Xiao W, Zhou Z, Xu Y, Pan Y, et al. FMDNet: Spatial‐frequency feature routing for low‐dose CT denoising. Journal of Applied Clinical Medical Physics. 2026. doi:10.1002/acm2.70656

BibTeX
@article{Yao2026,
  author = {Yao, Yujie and Liang, Yunsen and Xiao, Wen and Zhou, Zibo and Xu, Yaoxue and Pan, Yonghao and Xia, Xiao},
  title = {FMDNet: Spatial‐frequency feature routing for low‐dose CT denoising},
  journal = {Journal of Applied Clinical Medical Physics},
  year = {2026},
  doi = {10.1002/acm2.70656}
}
Kostenlos starten →