Data and similar resources.

Trained networks

We made the following already trained networks designed to specific tasks freely available together with example and training data.

Cytosol and nucleus 2D segmentation  

StarDist 2D segmentation of nucleus (DAPI) and U-net segmentation of cytoplasm (phalloidin staining).

Input data for cell segmentation and reference masks and edge weights used for training.

Download training data and networks:

  •  images.zip
  •  masks.zip
  •  cyto 
  •  nucleus

Format of input file is a 16-bit raw TIFF with 3 channels (i..e. POI, phalloidin, DAPI).
Run the prediction by:

conda create -n cytnuc
python


By default the first channel is channel of interest (e.g. staining), the second channel is phalloidin and the third channel is DAPI.
The output is a CSV file with the following structure:

Text

Microtubuli individual segmentation 3D

Microtubuli labeled by GFP in U2-OS cells.

Paramecium cilia 3D segmentation

Cilia of P.tetraulia with monoclonal antibody targeting tubulin (clone: AXO49).

40x epifluorescence z-stacks as input (non-confocal).

Segmentation masks created by Trupthi Josh
Paramecium 2D segmentation and tracking
Low-resolution 3D segmentation of single-molecules
High-resolution 3D segmentation of single-molecules