获取途径

作者

Yunlong Gao Xingyu Xie Hongmei Tang Hang Wang Xiaorui Wu Yi An Xiaohua Zhu Zhaoping Cheng Jiehua Xu Hairong Zheng Dong Liang Biao Li Hanzhong Wang Zhanli Hu

作者单位

²University of Chinese Academy of Sciences, Beijing, 100049, China ⁷Department of Nuclear Medicine, Zhuhai People’s Hospital (Zhuhai Clinical Medical College of Jinan University, The Affiliated Hospital of Beijing Institute of Technology), Zhuhai, China ⁵Department of Nuclear Medicine & Institute for Medical Imaging Technology, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China ⁴Department of Nuclear Medicine, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, 430000, China ¹Research Center for Medical AI, Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, Shenzhen, 518055, China ³Yunlong Gao and Xingyu Xie contributed equally to this work. ⁸Key Laboratory of Biomedical Imaging Science and System, Chinese Academy of Sciences, State Key Laboratory of Biomedical Imaging Science and System, Shenzhen, 518055, China ⁶Department of PET/CT, The First Affiliated Hospital of Shandong First Medical University, Jinan, China

刊名

Medical Image Analysis

年份

2026

卷号

Vol.114

页码

104265

ISSN

1361-8415

关键词

Diffusion model Autoregressive modeling Low-count PET PET image enhancement

摘要

Low-count positron emission tomography reduces injected activity or acquisition time, but fewer detected coincidence events compromise image quality and quantitative reliability. Existing diffusion-based PET enhancement methods commonly use generic, count-agnostic Gaussian schedules that do not explicitly represent acquisition/count-dependent variation in degradation severity between paired standard- and low-count reconstructions. Direct incorporation of multi-step history may also complicate M...更多

Low-count positron emission tomography reduces injected activity or acquisition time, but fewer detected coincidence events compromise image quality and quantitative reliability. Existing diffusion-based PET enhancement methods commonly use generic, count-agnostic Gaussian schedules that do not explicitly represent acquisition/count-dependent variation in degradation severity between paired standard- and low-count reconstructions. Direct incorporation of multi-step history may also complicate Markovian reverse inference. We propose a count-informed, endpoint-conditioned diffusion bridge in the reconstructed-image domain with gated autoregressive inference . The bridge is anchored to paired reconstructed endpoints: its conditional mean follows the standard-to-low-count residual, while a normalized expected-count trajectory controls progression along that residual and the aggregate latent variance. Measurement-level Poisson counting statistics motivate this trajectory from acquisition duration or injected-activity ratio, but no Poisson likelihood is imposed on reconstructed PET voxels. During reverse inference, GAI summarizes previous reverse states within an augmented state that retains a first-order Markov formulation and permits closed-form, count-conditioned updates. We evaluated the method on four-center total-body [F]FDG short-duration PET datasets and a simulated BrainWeb low-dose dataset. On the Shanghai Ruijin cohort, it increased PSNR by 5.33 dB and SSIM by 0.043 and reduced RMSE by 53.4% for 3 s acquisitions relative to the unenhanced short-duration PET input. At 1 s, PSNR increased by 5.91 dB and RMSE decreased by more than 60%. BrainWeb provided proof-of-concept evidence under controlled count reduction. These results support count-informed degradation modeling for low-count PET enhancement in the reconstructed-image domain.收起

文献类型

期刊