Synthetic Biventricular Heart Meshes Generated by the InSilicoHeartGen Pipeline for Electromechanical Simulations

-------------------------------------------------------------------------------------------------------------------------------------- If you use this dataset, please cite:Doste et al. (2026), Computer Methods and Programs in Biomedicine. https://doi.org/10.1016/j.cmpb.2026.109290 -------------------------------------------------------------------------------------------------------------------------------------- We present a database of synthetic biventricular heart meshes and associated simulation files, generated using the InSilicoHeartGen pipeline (Doste et al. 2026). This dataset comprises 50 cases representing adult human cardiac anatomy. The meshes were created from point clouds derived using a variational autoencoder, as described by Beetz et al. (2023) and Beetz, Banerjee, and Grau (2022). This generative approach produces anatomies that capture realistic population variability and preserve subpopulation-specific traits, consistent with the original dataset, in this case, the UK Biobank. From these point clouds, surface meshes were reconstructed using a ball pivoting algorithm, and the InSilicoHeartGen pipeline was then applied to generate volumetric meshes and simulation data suitable for in silico electrophysiological and electromechanical simulations. Cardiac Meshes:Each case contains: Folder with the coarse mesh, including: an .ensi file with the tetrahedral coarse mesh and associated field data, a .mat file containing all generated variables a folder with .csv files used for EP digital twinning as described in (Camps et al., 2025) Folder with the fine mesh, including: an .ensi file with the tetrahedral fine mesh and associated field data, a .mat file containing all generated variables a folder with .csv files used for EP digital twinning as described in (Camps et al., 2025) The triangular surface mesh with assigned labels 1 - epicardium2 - endocardium of the left ventricle (LV)3 - endocardium of the right ventricle (RV)4 - basal surface (only in cut geometries)5 - apex (in cut geometries)6 - septal endocardial wall of the RV9 - pulmonary valve10 - aortic valve12 - LV apex on the epicardium13 - mitral valve14 - tricuspid valve18 - RV apex on the endocardium19 - pulmonary valve on the epicardium (in closed meshes)20 - aortic valve on the epicardium (in closed meshes)23 - mitral valve on the epicardium (in closed meshes)24 - tricuspid valve on the epicardium (in closed meshes) A MonoAlg3D folder, containing: an hexahedral .ensi mesh with all generated fields, an .alg file formatted for running electrophysiological simulations in MonoAlg3D GPU solver (Sachetto Oliveira et al. 2018) automatically generated electrode positions. Some of the fields included in each mesh are: Fibre orientation (longitudinal, sheet, and normal directions) Transmurality Transmurality considering the RV septal endocardium as epicardial tissue Cobiveco coordinates Repolarisation gradients AHA segmentation Material tag definitions