Biventricular Heart Statistical Shape Model for Simulations

An adapted biventricular statistical shape model from Bai et al. (2015) provides a basis for computer simulations of cardiac electrophysiology and mechanics. The dataset includes two detailed meshes of the mean heart shape: a surface triangle mesh with 100k nodes and 200k elements, and a volumetric tetrahedral mesh with 479k nodes and 2.555 million elements. It also contains 100 principal components and variances for shape variation, plus 100 quasi-random shape instances generated by sampling weights within ±3 standard deviations.

Use Cases

• Simulating cardiac electrophysiology based on a detailed tetrahedral mesh of the myocardium.
• Performing cardiac mechanics simulations using the labeled surface regions (base, epicardium, LV/RV endocardium, apex).
• Generating synthetic heart shape variations for model testing by sampling from the 100 principal components.
• Analyzing anatomical shape variability using the provided principal components and variances.

Strengths

• Includes high-resolution meshes with specific node and element counts (e.g., 479k nodes, 2.555M tetrahedral elements).
• Provides 100 principal components for detailed statistical shape analysis and generation of 100 quasi-random instances.
• Mesh regions are explicitly labeled (base, epicardium, LV/RV endocardium, apex), facilitating targeted simulations.

Limitations

• Key metadata such as dataset size, row count, and last update date are unspecified across all sources.
• The model does not include any variations in right ventricular wall thickness, which may limit mechanical realism.
• Column names and a complete data schema are not provided, making precise data structure inference necessary.

Provenance

Source

Adapted from the biventricular statistical shape model by Bai et al. (2015); author is Steffen Schuler, associated with Karlsruhe Institute of Technology.

Collection Method

Original disconnected surfaces were processed (clipped, merged, remeshed) using Instant Meshes and Gmsh; principal components were Laplacian interpolated.