Published February 19, 2020 | Version v2
Other Open

Platynereis EM network weights

  • 1. EMBL Heidelberg

Description

Weights for trained Convolutional Neural Networks used in the publication Whole-body integration of gene expression and single-cell morphology. We provide weights for 3d U-Nets segmenting the following structures:
- cell membranes
- cilia
- cuticle
- nuclei

For details on how to use these networks were trained, see https://github.com/platybrowser/platybrowser-backend/tree/master/segmentation. The training data is available at https://doi.org/10.5281/zenodo.3675220.

This version provides the state dicts rather than pickled models, which are provided in the first version.

Files

Files (232.6 MB)

Name Size Download all
md5:c7027b44cbaa334105248397c43de6fe
75.5 MB Download
md5:d3660de597b8f56c74ce73178f4f3bd8
39.3 MB Download
md5:1fa34bcab44c56822684d5ba38cad0d4
39.3 MB Download
md5:21ecc7e77cbec219d5a0cf3e1b0c24b0
39.3 MB Download
md5:42d9c4805dcdb93c1a6e7fe765966a4f
39.3 MB Download