Make sure to have a running napari installation.
In the environment with your napari installation run:
$ pip install merge-stardist-masks
Within napari go to
Plugins -> Install/Uninstall Plugins... and search for
napari-merge-stardist-masks in the lower list. Then click on the blue install
Make sure to restart napari after the installation. If you do not find the plugins, go
Plugins -> Install/Uninstall Plugins... and toggle the checkboxes in the upper list for
stardist-napari and napari-merge-stardist-masks.
Download one of the pre-trained StarDist models from here and unzip the file.
Run a segmentation#
Load sample data with
File -> Open sample -> StarDist OPP sample data
Plugins -> StarDist OPP. Two widgets will open, the StarDist plugin and this plugin.
All the parameters in the StarDist plugin should be correctly set already. Make sure that the axes in the field
Image Axesare correct, for a 3D image it should be
Custom 2D/3Din the field
Model Typeand choose the directory where you unzipped the pre-trained model in the
Custom Modelfield. See the image below for the correct settings.
Run. And wait until the CNN calculates the outputs. The outputs of the CNN are displayed once they are calculated.
In the StarDist OPP widget, select again the path to the unzipped pre-trained model in the field
model path. Then select
StarDist distances (data)and
StarDist probability (data)for the
You can play around with the other fields. However, this might lead to errors. For 3D images, you should set
1.00the other settings are already fine. See the following image for proper settings.
Runin the StarDist OPP widget. The post-processing starts and might take some time (on our machine it takes ~10 minutes). Once the post-processing is done, the label image will be shown in the viewer.