HorusEye: a self-supervised foundation model for generalizable X-ray tomography restoration
X-ray tomography is widely used across scientific and clinical domains, yet image degradation remains a major obstacle to reliable analysis, particularly under low-dose or data-scarce conditions. Existing restoration methods are typically designed for specific modalities and predefined degradation, limiting their generalizability. Here we show that image restoration can instead be formulated as learning realistic, nonparametric acquisition degradation processes directly from data. We introduce HorusEye, a self-supervised foundation model for X-ray tomography restoration that leverages interslice contrastive pretraining to jointly learn structural priors and degradation without paired supervision or predefined assumptions. Trained on over 100 million images, HorusEye generalizes across diverse modalities, restoration tasks and previously unseen imaging modalities, consistently outperforming task-specific approaches. Extensive evaluations demonstrate improved photon efficiency and recovery of high-frequency information. Clinical studies further demonstrate enhanced detectability of low-contrast anatomy and lesions, as well as improved performance on downstream tasks, highlighting HorusEye as a general postprocessing tool for X-ray tomography.
Authors
- Longxi Zhou (ORCID: https://orcid.org/0000-0003-0116-9361)
- Dexuan Xie (ORCID: https://orcid.org/0000-0002-7072-6275)
- X.-J. Gao
- Zhaowen Qiu (ORCID: https://orcid.org/0000-0001-5292-0333)
- Changchun Yang (ORCID: https://orcid.org/0000-0001-9567-630X)
- Ricardo Henao (ORCID: https://orcid.org/0000-0003-4980-845X)
- Gianluca Setti (ORCID: https://orcid.org/0000-0002-2496-1856)
- Zhongyi Han (ORCID: https://orcid.org/0000-0003-2851-193X)
- Kai Kang (ORCID: https://orcid.org/0000-0001-5758-5570)
- Yuetan Chu (ORCID: https://orcid.org/0000-0003-1729-831X)
- Suyu Dong
- Lianming Wu
- Yuan Cheng
- Xin Liu
- Shaodong Cao
- Xianglin Meng
- Xin Guo
- Yuan Qi
- Gongning Luo
- Yue Li
- Xigang Xiao
Institutions
- University College of the Cayman Islands (KY)
- Shanghai Academy of Social Sciences (CN)
- Harbin Medical University (CN)
- Shanghai Jiao Tong University (CN)
- Harbin Institute of Technology (CN)
- Renji Hospital (CN)
- Southern University of Science and Technology (CN)
- Heilongjiang International University (CN)
- Truman Bodden Law School (KY)
- Fourth Affiliated Hospital of Harbin Medical University (CN)
- Shanghai Academy of Environmental Sciences (CN)
- First Affiliated Hospital of Harbin Medical University (CN)
- Northeast Forestry University (CN)
- King Abdullah University of Science and Technology (SA)
Publication Details
- Journal
- Nature Computational Science
- Published
- 2026-03-27
- DOI
- https://doi.org/10.1038/s43588-026-00973-3
- Primary Topic
- Medical Imaging Techniques and Applications
- Type
- article
- Field-Weighted Citation Impact
- 0.00