VesselVAE: Recursive Variational Autoencoders for 3D Blood Vessel Synthesis

We present a data-driven generative framework for synthesizing blood vessel 3D geometry. This is a challenging task due to the complexity of vascular systems, which are highly variating in shape, size, and structure. Existing model-based methods provide some degree of control and variation in the st...

Descripción completa

Guardado en:
Detalles Bibliográficos
Autores principales: Feldman, Paula, Fainstein, Miguel, Siless, Viviana, Delrieux, Claudio, Iarussi, Emmanuel
Formato: Objeto de conferencia
Lenguaje:Inglés
Publicado: 2024
Materias:
Ciencias Informáticas
Vascular 3D model
Generative modeling
Neural Networks
Acceso en línea:http://sedici.unlp.edu.ar/handle/10915/178481
Aporte de:
SEDICI (UNLP) de Universidad Nacional de La Plata
id I19-R120-10915-178481
record_format dspace
spelling I19-R120-10915-1784812025-04-25T20:07:47Z http://sedici.unlp.edu.ar/handle/10915/178481 VesselVAE: Recursive Variational Autoencoders for 3D Blood Vessel Synthesis Feldman, Paula Fainstein, Miguel Siless, Viviana Delrieux, Claudio Iarussi, Emmanuel 2024-08 2024 2025-04-25T14:32:56Z en Ciencias Informáticas Vascular 3D model Generative modeling Neural Networks We present a data-driven generative framework for synthesizing blood vessel 3D geometry. This is a challenging task due to the complexity of vascular systems, which are highly variating in shape, size, and structure. Existing model-based methods provide some degree of control and variation in the structures produced, but fail to capture the diversity of actual anatomical data. We developed VesselVAE, a recursive variational Neural Network that fully exploits the hierarchical organization of the vessel and learns a low-dimensional manifold encoding branch connectivity along with geometry features describing the target surface. After training, the VesselVAE latent space can be sampled to generate new vessel geometries. To the best of our knowledge, this work is the first to utilize this technique for synthesizing blood vessels. We achieve similarities of synthetic and real data for radius (.97), length (.95), and tortuosity (.96). By leveraging the power of deep neural networks, we generate 3D models of blood vessels that are both accurate and diverse, which is crucial for medical and surgical training, hemodynamic simulations, and many other purposes. Sociedad Argentina de Informática e Investigación Operativa Objeto de conferencia Objeto de conferencia http://creativecommons.org/licenses/by-nc-sa/4.0/ Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0) application/pdf 60-69
institution Universidad Nacional de La Plata
institution_str I-19
repository_str R-120
collection SEDICI (UNLP)
language Inglés
topic Ciencias Informáticas
Vascular 3D model
Generative modeling
Neural Networks
spellingShingle Ciencias Informáticas
Vascular 3D model
Generative modeling
Neural Networks
Feldman, Paula
Fainstein, Miguel
Siless, Viviana
Delrieux, Claudio
Iarussi, Emmanuel
VesselVAE: Recursive Variational Autoencoders for 3D Blood Vessel Synthesis
topic_facet Ciencias Informáticas
Vascular 3D model
Generative modeling
Neural Networks
description We present a data-driven generative framework for synthesizing blood vessel 3D geometry. This is a challenging task due to the complexity of vascular systems, which are highly variating in shape, size, and structure. Existing model-based methods provide some degree of control and variation in the structures produced, but fail to capture the diversity of actual anatomical data. We developed VesselVAE, a recursive variational Neural Network that fully exploits the hierarchical organization of the vessel and learns a low-dimensional manifold encoding branch connectivity along with geometry features describing the target surface. After training, the VesselVAE latent space can be sampled to generate new vessel geometries. To the best of our knowledge, this work is the first to utilize this technique for synthesizing blood vessels. We achieve similarities of synthetic and real data for radius (.97), length (.95), and tortuosity (.96). By leveraging the power of deep neural networks, we generate 3D models of blood vessels that are both accurate and diverse, which is crucial for medical and surgical training, hemodynamic simulations, and many other purposes.
format Objeto de conferencia
Objeto de conferencia
author Feldman, Paula
Fainstein, Miguel
Siless, Viviana
Delrieux, Claudio
Iarussi, Emmanuel
author_facet Feldman, Paula
Fainstein, Miguel
Siless, Viviana
Delrieux, Claudio
Iarussi, Emmanuel
author_sort Feldman, Paula
title VesselVAE: Recursive Variational Autoencoders for 3D Blood Vessel Synthesis
title_short VesselVAE: Recursive Variational Autoencoders for 3D Blood Vessel Synthesis
title_full VesselVAE: Recursive Variational Autoencoders for 3D Blood Vessel Synthesis
title_fullStr VesselVAE: Recursive Variational Autoencoders for 3D Blood Vessel Synthesis
title_full_unstemmed VesselVAE: Recursive Variational Autoencoders for 3D Blood Vessel Synthesis
title_sort vesselvae: recursive variational autoencoders for 3d blood vessel synthesis
publishDate 2024
url http://sedici.unlp.edu.ar/handle/10915/178481
work_keys_str_mv AT feldmanpaula vesselvaerecursivevariationalautoencodersfor3dbloodvesselsynthesis
AT fainsteinmiguel vesselvaerecursivevariationalautoencodersfor3dbloodvesselsynthesis
AT silessviviana vesselvaerecursivevariationalautoencodersfor3dbloodvesselsynthesis
AT delrieuxclaudio vesselvaerecursivevariationalautoencodersfor3dbloodvesselsynthesis
AT iarussiemmanuel vesselvaerecursivevariationalautoencodersfor3dbloodvesselsynthesis
_version_ 1847925373923229696

Ejemplares similares