Visual language model-assisted CT denoising via text-guided diffusion and fidelity maintenance
Journal of X-Ray Science and Technology · 2026-03 · 34 (2) · 151-169
DOI: 10.1177/08953996251372739 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of X-Ray Science and Technology, online erschienen am 12.03.2026. Verfasst von Ye Shen und 8 weiteren Autor:innen. Beteiligte Einrichtung: Department of Henan Key Laboratory of Imaging and Intelligent Processing, Information Engineering University, Zhengzhou, China. Hauptthemen laut MEDLINE: Tomography, X-Ray Computed, Image Processing, Computer-Assisted. Die Arbeit nennt 48 Literaturangaben. Gefördert von: Innovative Research Group Project of the National Natural Science Foundation of China, the 2023 Central Plains Science and Technology Innovation Leading Talents Project “Research on Key Technologies of Linear Trajectory Intelligent 3D CT System”, the Natural Science Foundation of Henan.
Abstract (Originalsprache)
Reducing radiation dose in computed tomography (CT) and photon-counting CT (PCCT) is crucial for patient safety, but lower doses introduce noise that degrades image quality. Existing denoising methods often rely on supervised learning of paired data or are based on specific noise assumptions, which poses challenges in clinical practice. A novel Visual-Language Model-assisted CT Denoising (VLD) framework is proposed to address CT image noise while preserving diagnostic fidelity through semantic guidance. Our method innovatively leverages the human-level knowledge embedded in multimodal visual-language models and applies it to the field of CT image denoising. This approach enables the diffusion model to perform restoration guided by semantic understanding. Meanwhile, a tri-domain consistency framework has been proposed to further enhance image quality by progressively refining details while preserving structural integrity. Extensive experiments on both simulated CT and real PCCT data demonstrate that the VLD method generates high-quality reconstruction images and exhibits robust generalization to new scenarios. In simulation experiments, the VLD method achieves average improvements of 0.95 dB and 1.21 dB in peak signal-to-noise ratio under the 5000-photon number condition, outperforming the WGAN and FBPConvNet methods, which require paired data.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Department of Henan Key Laboratory of Imaging and Intelligent Processing, Information Engineering University, Zhengzhou, China.
ORCID-Profile: Ye Shen, Ailong Cai, Xinrui Zhang, Zhizhong Zheng
Förderung
- Innovative Research Group Project of the National Natural Science Foundation of China
- the 2023 Central Plains Science and Technology Innovation Leading Talents Project “Research on Key Technologies of Linear Trajectory Intelligent 3D CT System”
- the Natural Science Foundation of Henan
Zitieren
APA
Shen, Y., Liang, N., Cai, A., Zhang, X., Wang, Y., Ren, J., Zheng, Z., Li, L., & Yan, B. (2026). Visual language model-assisted CT denoising via text-guided diffusion and fidelity maintenance. Journal of X-Ray Science and Technology, 34(2), 151-169. https://doi.org/10.1177/08953996251372739
Vancouver
Shen Y, Liang N, Cai A, Zhang X, Wang Y, Ren J, et al. Visual language model-assisted CT denoising via text-guided diffusion and fidelity maintenance. Journal of X-Ray Science and Technology. 2026;34(2):151-169. doi:10.1177/08953996251372739
BibTeX
@article{Shen2026,
author = {Shen, Ye and Liang, Ningning and Cai, Ailong and Zhang, Xinrui and Wang, Yizhong and Ren, Junru and Zheng, Zhizhong and Li, Lei and Yan, Bin},
title = {Visual language model-assisted CT denoising via text-guided diffusion and fidelity maintenance},
journal = {Journal of X-Ray Science and Technology},
year = {2026},
volume = {34},
number = {2},
pages = {151--169},
doi = {10.1177/08953996251372739}
}
EUVIMED