VesselGPT: Autoregressive Modeling of Vascular Geometry

Versión final publicada: Feldman, P., Sinnona, M., Delrieux, C., Siless, V., Iarussi, E. (2026). VesselGPT: Autoregressive Modeling of Vascular Geometry. In: Gee, J.C., et al. Medical Image Computing and Computer Assisted Intervention – MICCAI 2025. MICCAI 2025. Lecture Notes in Computer Science,...

Descripción completa

Guardado en:
Detalles Bibliográficos
Autores principales: Feldman, Paula, Sinnona, Martín, Delrieux, Claudio, Siless, Viviana, Iarussi, Emmanuel
Formato: info:eu-repo/semantics/Preprint
Lenguaje:Inglés
Publicado: 2025
Materias:
Inteligencia Artificial
Artificial Intelligence
Tratamiento médico
Medical Treatment
Vasos Sanguíneos
Blood Vessels
Acceso en línea:https://doi.org/10.48550/arXiv.2505.13318
https://repositorio.utdt.edu/handle/20.500.13098/13402
Aporte de:
Repositorio Digital Universidad Torcuato Di Tella de Universidad Torcuato Di Tella
id I57-R163-20.500.13098-13402
record_format dspace
spelling I57-R163-20.500.13098-134022025-10-23T11:54:41Z VesselGPT: Autoregressive Modeling of Vascular Geometry Feldman, Paula Sinnona, Martín Delrieux, Claudio Siless, Viviana Iarussi, Emmanuel Inteligencia Artificial Artificial Intelligence Tratamiento médico Medical Treatment Vasos Sanguíneos Blood Vessels Versión final publicada: Feldman, P., Sinnona, M., Delrieux, C., Siless, V., Iarussi, E. (2026). VesselGPT: Autoregressive Modeling of Vascular Geometry. In: Gee, J.C., et al. Medical Image Computing and Computer Assisted Intervention – MICCAI 2025. MICCAI 2025. Lecture Notes in Computer Science, vol 15975. Springer, Cham. https://doi.org/10.1007/978-3-032-05325-1_63 Anatomical trees are critical for clinical diagnosis and treatment planning, yet their complex and diverse geometry make accurate representation a significant challenge. Motivated by the latest advances in large language models, we introduce an autoregressive method for synthesizing anatomical trees. Our approach first embeds vessel structures into a learned discrete vocabulary using a VQ-VAE architecture, then models their generation autoregressively with a GPT-2 model. This method effectively captures intricate geometries and branching patterns, enabling realistic vascular tree synthesis. Comprehensive qualitative and quantitative evaluations reveal that our technique achieves high-fidelity tree reconstruction with compact discrete representations. Moreover, our B-spline representation of vessel cross-sections preserves critical morphological details that are often overlooked in previous’ methods parameterizations. To the best of our knowledge, this work is the first to generate blood vessels in an autoregressive manner. Code, data, and trained models will be made available. Feldman, P., et al. (2025). VesselGPT: Autoregressive Modeling of Vascular Geometry. Arxiv. https://doi.org/10.48550/arXiv.2505.13318 2025-05-27T18:05:39Z 2025-05-19 info:eu-repo/semantics/Preprint https://doi.org/10.48550/arXiv.2505.13318 https://repositorio.utdt.edu/handle/20.500.13098/13402 eng Arxiv info:eu-repo/semantics/openAccess https://creativecommons.org/licenses/by/4.0/deed.es 11 p. application/pdf application/pdf
institution Universidad Torcuato Di Tella
institution_str I-57
repository_str R-163
collection Repositorio Digital Universidad Torcuato Di Tella
language Inglés
orig_language_str_mv eng
topic Inteligencia Artificial
Artificial Intelligence
Tratamiento médico
Medical Treatment
Vasos Sanguíneos
Blood Vessels
spellingShingle Inteligencia Artificial
Artificial Intelligence
Tratamiento médico
Medical Treatment
Vasos Sanguíneos
Blood Vessels
Feldman, Paula
Sinnona, Martín
Delrieux, Claudio
Siless, Viviana
Iarussi, Emmanuel
VesselGPT: Autoregressive Modeling of Vascular Geometry
topic_facet Inteligencia Artificial
Artificial Intelligence
Tratamiento médico
Medical Treatment
Vasos Sanguíneos
Blood Vessels
description Versión final publicada: Feldman, P., Sinnona, M., Delrieux, C., Siless, V., Iarussi, E. (2026). VesselGPT: Autoregressive Modeling of Vascular Geometry. In: Gee, J.C., et al. Medical Image Computing and Computer Assisted Intervention – MICCAI 2025. MICCAI 2025. Lecture Notes in Computer Science, vol 15975. Springer, Cham. https://doi.org/10.1007/978-3-032-05325-1_63
format info:eu-repo/semantics/Preprint
author Feldman, Paula
Sinnona, Martín
Delrieux, Claudio
Siless, Viviana
Iarussi, Emmanuel
author_facet Feldman, Paula
Sinnona, Martín
Delrieux, Claudio
Siless, Viviana
Iarussi, Emmanuel
author_sort Feldman, Paula
title VesselGPT: Autoregressive Modeling of Vascular Geometry
title_short VesselGPT: Autoregressive Modeling of Vascular Geometry
title_full VesselGPT: Autoregressive Modeling of Vascular Geometry
title_fullStr VesselGPT: Autoregressive Modeling of Vascular Geometry
title_full_unstemmed VesselGPT: Autoregressive Modeling of Vascular Geometry
title_sort vesselgpt: autoregressive modeling of vascular geometry
publishDate 2025
url https://doi.org/10.48550/arXiv.2505.13318
https://repositorio.utdt.edu/handle/20.500.13098/13402
work_keys_str_mv AT feldmanpaula vesselgptautoregressivemodelingofvasculargeometry
AT sinnonamartin vesselgptautoregressivemodelingofvasculargeometry
AT delrieuxclaudio vesselgptautoregressivemodelingofvasculargeometry
AT silessviviana vesselgptautoregressivemodelingofvasculargeometry
AT iarussiemmanuel vesselgptautoregressivemodelingofvasculargeometry
_version_ 1847744043872681984

Ejemplares similares