EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

SwinPix: A Swin Transformer-based Pix2Pix Framework for Low-Dose PET Denoising Using Multi-level Inputs Toward Standard-Dose Quality

Mohammad Saber Azimi, Vahid Felfelian, Habibollah Dadgar, Hossein Arabi, Habib Zaidi

Journal of Imaging Informatics in Medicine · 09.03.2026

DOI: 10.1007/s10278-026-01889-0 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Imaging Informatics in Medicine, online erschienen am 09.03.2026. Verfasst von Mohammad Saber Azimi und 4 weiteren Autor:innen. Beteiligte Einrichtung: Doctoral School of Applied Informatics and Applied Mathematics, Óbuda University, Budapest, Hungary. Die Arbeit nennt 52 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung, Óbuda University.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

This study introduces SwinPix, a novel network architecture designed to explore the effectiveness of multi-level low-dose (LD) PET inputs as prior knowledge for standard-dose (SD) PET image prediction. By employing SwinPix architecture, we assess the performance of single-input and multi-input models trained with PET data at 4%, 6%, and 10% dose levels. Two models were developed: the first was trained using a single input corresponding to 4%, 6%, and 10% LD PET images, while the second considered a multi-input approach, utilizing three lower-dose inputs to predict the corresponding SD PET images. The performance of the six models was evaluated using structural similarity index (SSIM), peak signal-to-noise ratio (PSNR), mean standardized uptake value (SUV mean ) bias, SUV max bias, and root mean square error (RMSE) within the entire head region and malignant lesions. The SwinPix multi-input model outperformed single-input versions across all dose levels. At 4%, PSNR increased by 13%, SSIM improved from 0.97 to 0.99, while RMSE (lesion/head) dropped by 82–86%. Similarly, SUV mean and SUV max biases decreased by 78% and 58%, respectively. At 6% and 10% dose levels, SwinPix showed comparable improvements, reducing RMSE by over 40% and SUV biases by up to 53%. Compared to Pix2Pix and Swin Transformer, SwinPix consistently achieved the best reconstruction quality. All improvements were statistically significant ( p < 0.01), supporting the effectiveness of multi-input SwinPix for accurate LD PET imaging. SwinPix offers a hybrid transformer-based solution for LD PET reconstruction. By incorporating multi-level LD PET inputs through a Generative adversarial network (GAN) framework, it enhances image quality and lesion quantification while maintaining computational efficiency, supporting its potential for clinical deployment.

Schlagwörter

Stichwörter der Autor:innen

Deep learningLow-dose PETMulti-input reconstructionPET denoisingStandard-dose PETSwin Transformer

Autor:innen und Institutionen

  • Doctoral School of Applied Informatics and Applied Mathematics, Óbuda University, Budapest, Hungary.
  • University Research and Innovation Center, Óbudabuda University, Budapest, Hungary.
  • Department of Medical Radiation Engineering, Shahid Beheshti University, Tehran, Iran.
  • Cancer Research Center, RAZAVI Hospital, Imam Reza International University, Mashhad, Iran.
  • Division of Nuclear Medicine & Molecular Imaging, Geneva University Hospital, CH-1211, Geneva, Switzerland.
  • University Research and Innovation Center, Óbudabuda University, Budapest, Hungary. habib.zaidi@hcuge.ch.

Förderung

  • Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung
  • Óbuda University

Zitieren

APA

Azimi, M. S., Felfelian, V., Dadgar, H., Arabi, H., & Zaidi, H. (2026). SwinPix: A Swin Transformer-based Pix2Pix Framework for Low-Dose PET Denoising Using Multi-level Inputs Toward Standard-Dose Quality. Journal of Imaging Informatics in Medicine. https://doi.org/10.1007/s10278-026-01889-0

Vancouver

Azimi MS, Felfelian V, Dadgar H, Arabi H, Zaidi H. SwinPix: A Swin Transformer-based Pix2Pix Framework for Low-Dose PET Denoising Using Multi-level Inputs Toward Standard-Dose Quality. Journal of Imaging Informatics in Medicine. 2026. doi:10.1007/s10278-026-01889-0

BibTeX
@article{Azimi2026,
  author = {Azimi, Mohammad Saber and Felfelian, Vahid and Dadgar, Habibollah and Arabi, Hossein and Zaidi, Habib},
  title = {SwinPix: A Swin Transformer-based Pix2Pix Framework for Low-Dose PET Denoising Using Multi-level Inputs Toward Standard-Dose Quality},
  journal = {Journal of Imaging Informatics in Medicine},
  year = {2026},
  doi = {10.1007/s10278-026-01889-0}
}
Kostenlos starten →