Metraipolis › archive › dc_dcf6ce3e42bc

TomoTransformer: Towards a Foundation Model for CT Reconstruction — arXiv:2609.37605

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.37605
License (as archived)
CC0-1.0
Archived
2026-10-07T17:33:01.974Z
SHA-256 of the archived text
2eda0b1fcace67dc262ffec0431d45deb4f3243a0178310f5bf39f489e136048
Stable URL of this page
https://metraipolis.com/documento/dc_dcf6ce3e42bc

How to cite

TomoTransformer: Towards a Foundation Model for CT Reconstruction — arXiv:2609.37605 — arXiv (en), https://arxiv.org/abs/2609.37605 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_dcf6ce3e42bc (Metraipolis archive dc_dcf6ce3e42bc, sha256 2eda0b1fcace67dc262ffec0431d45deb4f3243a0178310f5bf39f489e136048)
BibTeX
@misc{dcdcf6ce3e42bc,
  title = {TomoTransformer: Towards a Foundation Model for CT Reconstruction — arXiv:2609.37605},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.37605},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_dcf6ce3e42bc; sha256 2eda0b1fcace67dc262ffec0431d45deb4f3243a0178310f5bf39f489e136048},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.37605. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Supervised deep learning has advanced sparse-view tomographic reconstruction. However, conventional models, which typically map filtered back-projection (FBP) images or sinograms to clean reconstructions, are brittle under distribution shifts. Because they require retraining whenever projection counts and angles, detector resolutions, or data distributions change, their deployment in real-world applications remains limited. To address this, we introduce TomoTransformer, a transformer-based architecture that treats each \textit{local} filtered projection as an individual token and predicts missing views via self-attention. Crucially, TomoTransformer operates in a \emph{back-projection space} that separates projections across spatial locations, making view interpolation geometrically well-posed and invariant to detector size. This design yields a single foundation model that can process any number of input projections, at arbitrary angular locations and detector dimensions, and query any number of target angles without retraining. Trained on a large-scale dataset spanning diverse medical CT anatomies and natural images, TomoTransformer generalizes effectively across anatomies, materials, and resolutions. Extensive evaluations on several benchmark sparse-view datasets show that TomoTransformer significantly outperforms concurrent multi-purpose models like ViewTrans and matches or exceeds strong protocol-specific baselines, while remaining fully agnostic to the number of input and target projections. Furthermore, the model demonstrates robust zero-shot generalization on real experimental nanoscale brain data collected from an X-ray synchrotron, showcasing its practical utility for real-world applications.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-30 in cs.LG).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2609.37605 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-07T16:59:48.565Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-09-30?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).