There is a newer version of the record available.

Published April 30, 2023 | Version 1.0.0
Dataset Open

Simulation dataset to benchmark 3D force inference methods

Authors/Creators

  • 1. Collège de France

Description

Simulated dataset of 46 images (.tif), segmentation masks (.tif), corresponding to foam-like cell structures (early embryos) of various cell numbers (2 to 11), cell sizes and interfacial tensions.
The ground truth tensions and pressures to be inferred are provided as Numpy arrays (.npy).
Non-manifold multimaterial meshes corresponding meshes are also provided as binary files (.rec) and may be opened with our delaunay-watershed Python code, that can be installed via

pip install delaunay-watershed-3d

 

Files

BenchmarkDataset_for_3DForceInference.zip

Files (79.1 MB)

Name Size Download all
md5:9ac9344f350a2f2fc71c3830186555a0
79.1 MB Preview Download

Additional details

Funding

European Commission
DeepEmbryo - Reverse-engineering the development of embryos with physics-informed machine learning 949267

References

  • Ichbiah, Sacha, et al. (2023). Embryo mechanics cartography: inference of 3D force atlases from fluorescence microscopy. bioRxiv, 2023-04.