Detecting cells in 3D with cellfinder#

In this tutorial, you will use the cellfinder plugin for napari to detect cells in a 3D image.

The aim of this tutorial is to get cell detection up and running. The image we use here is very small and has some artefacts. The results are not very good, but it should be enough to see the plugin working! When using cellfinder on your own data, there are many parameters that can be adjusted, and cellfinder will need to be retrained for optimal performance. For more information, please see the full cellfinder documentation.

Note

You will need napari installed on your computer - please follow napari’s installation instructions to do so (including their recommendation to use a conda environment).

  1. Open napari.

  2. Install cellfinder by selecting Plugins > Install/Uninstall plugins and searching for cellfinder in the searchbox. If it is not installed yet, click on the Install button.

Caution

This may take a while due to the various dependencies being installed.

On Silicon Macs you may have to run conda install hdf5 on the command line (in your conda environment) first for the installation to be successful.

  1. Open the cell detection widget by selecting Plugins > cellfinder > Cell detection in the napari menu bar near the top left of the window. cellfinder detection widget

    The cell detection widget appears on the right-hand side of the window.

  2. Load some sample data File > Open sample > Sample data (cellfinder). This will open a small two-channel 3D image.

  3. Set the signal image to Signal

  4. Set the background image to Background

  5. Ensure that Voxel size (z) is set to 5.00, Voxel size (y) is set to 2.00, and Voxel size (x) is set to 2.00.

  6. Click Run (you may need to scroll down)

    Detected cells

    Detected cells are highlighted by yellow circles.

  7. Ensure that the Detected layer is highlighted

  8. Save the detected cells by clicking File > Save selected layer(s), choosing a directory and a filename ending in .xml (e.g. cells.xml).

Hint

For more information about how to use the detection plugin to detect cells in your own data, please see the full cell detection guide.