This repository contains configuration files used to reproduce the segmentation experiments from the paper An ex vivo system to study cellular dynamics underlying mouse peri-implantation development:
@article{ICHIKAWA2022,
title = {An ex vivo system to study cellular dynamics underlying mouse peri-implantation development},
journal = {Developmental Cell},
year = {2022},
issn = {1534-5807},
doi = {https://doi.org/10.1016/j.devcel.2021.12.023},
url = {https://www.sciencedirect.com/science/article/pii/S1534580721010431},
author = {Takafumi Ichikawa and Hui Ting Zhang and Laura Panavaite and Anna Erzberger and Dimitri Fabrèges and Rene Snajder and Adrian Wolny and Ekaterina Korotkevich and Nobuko Tsuchida-Straeten and Lars Hufnagel and Anna Kreshuk and Takashi Hiiragi},
keywords = {mouse embryonic development, embryo implantation, egg cylinder formation, epiblast morphogenesis, lumen formation, tissue-tissue interaction, mechano-chemical interplay, embryo culture, live-imaging, quantitative image analysis}
}
Neural networks used to predict the embryo cell boundaries were trained with pytorch-3dunet (see configuration files in unet_configs).
Given the cell boundary network the final segmentation results were computed using plant-seg. Relevant plant-seg configuration files can be found in plantseg_configs.
For networks trained with sparsely annotated embryo cells we used SPOCO. SPOCO configuration files can be found in spoco_configs.