Installation

ewoksndreg provides functionality to register images using ewoks tasks and workflows.

The optimal way to use the tasks is through orange widgets, as visualizing the images is important to assess success of the alignment.

Step 1 - Environment

We assume that your machine has python installed. You can either set up a new virtual environment to install this package in or use an existing one. If you are using an existing environment you can skip the creation of the environment.

Go to the folder where you want to create the environment and create a python environment with (.venv an be replaced by a fitting name for the environment)

python -m venv .venv

you can then activate the virtual environment with

source .venv/bin/activate

This setup is valid for Unix/macOS. For a more thorough walkthrough and the windows version, see the python guide

Step 2 - Package

We can use pip to install the ewoksndreg package with

pip install ewoksndreg[full]

The environment must be activated for this, which is indicated by the (venv) in the beginning of the line in the terminal

This will install the package and all its dependencies. If you have issues with memory space and don’t need all options for backends, an alternative would be to only install the following packages as this leaves out Kornia and its dependency pytorch which are fairly big.

pip install ewoksndreg
pip install simpleitk
pip install scikit-image

Step 3 - Orange

It should now be possible to use the tasks in ewoks workflows. If we want to use the GUIs for these tasks, which allow you to interactivly do the registration, we also need to install orange.

pip install orange3

We can now launch the GUI with

orange-canvas

Step 4 - First workflow

In Orange we can first execute an example workflow, to see if everything works correctly. The window with example workflows can be found through the example tab in the ‘Welcome to Orange’ Window or through > Help > Example Workflows. You can open either of the workflows ‘Feature-based Image Registration’ and ‘Intensity-based Image Registration’.

To trigger the workflow you can open the ‘2D Example Stack’ Task and press ‘Trigger’. After a short computation, you should find the aligned images in the ‘Align’ task

Step 5 - Next steps

For more information about the specific widgets and underlying tasks, see Available Widgets.

A small introduction and some links for further reading regarding the theory can be found in Theory.