Veröffentlicht 5. August 2025 | Version 1.2
Dataset Offen

SynthAorta Dataset

Beschreibung

SynthAorta Dataset

This repository contains the SynthAorta dataset, based on the paper "SynthAorta: A 3D Mesh Dataset of Parametrized Physiological Healthy Aortas" (arXiv:2409.08635). Any new updates or code related to the dataset will be published on the corresponding GitHub repository: https://github.com/domagoj-bosnjak/SynthAorta.

Dataset info

The dataset contains 30,000 synthetic geometries of physiological, healthy aortas, including three supraaortic vessels. Specifically, the dataset contains:

  • Structured hexahedral meshes; 4 refinement levels with 226, 1,792, 14,336, and 114,688 elements, respectively.
  • Consistent meshes, with the same node numbering for every example.
  • Centerlines with local radius data, enabling the recovery of the smooth surface using the convolution surface framework.
  • Code for outputting the meshes to the .msh file format, easily transferable to other FEM software.
  • The main parameter file, parameters_SynthAorta.csv, with all described geometric features. All accompanying files provide node-wise radii for different aortic regions. All files are aligned by case number, with consistent skeleton sampling across geometries. Full parameter definitions are available in the related publication.

Note: the dataset contains around 90,000 files in total.

Usage and file formats

All data is stored in binary format for efficiency. For more information, refer to the README.txt file. Minimal MATLAB/Octave code is also uploaded, which handles the input/output of meshes and skeletons, as well as elementary visualizations. The meshes are outputted in the standard .msh format, which can be parsed/converted by any modern FEM software or library. 

The examples folder contains extracted meshes for initial visualization, requiring the open-source software Gmsh.

Contact

All feedback is welcome, especially if you encounter problems or issues using the dataset. Please get in touch with us at bosnjak@tugraz.at or gmelito@tugraz.at.

 

Dateien

Dateien (26.2 GB)

Name Größe Alle herunterladen
md5:b61f5dbb54ac00368a421498a72d1769
26.2 GB Download

Weitere Details

Verknüpfte Arbeiten

Is derived from
Journal article: 10.48550/arXiv.2409.08635 (DOI)