dsblab/registrationtools repository overview

⁠Registrationtools docker

The dockerfile makes an image that contains a set of relevant packages for image registration in python.

The relevant packages for image registration that contains are:

• registrationtools⁠: Wrapper of blockmatching algorithm as implemented in vt.
• vt-python⁠: Original package over which this package makes a wrapper.
• SimpelITK⁠: Classical package for image registration, wrapper over the ITK package.
• skimage⁠: Well-stablished package for image manipulation in python based on numpy.

In addition to these packages, other packages are installed (e.g. numpy, matplotlib...). A list can be seen in environment.yaml.

Constructed images are already deposited in dsblab/registrationtools⁠, so there is not need of building the image except if you want to extend further capabilities.

⁠Running the image

⁠Interactive Shell

Start the container in interactive format.

docker run -it \
    --mount type=bind,source="$(pwd)",target=/home \
    dsblab/registrationtools:v0.1

Activate the conda image.

source activate registration
cd home

And now all the packages and modules will be loaded. You can start a python shell:

python

or execute packages installed from the command line (e.g. blockmatching)

blockmatching -h

⁠Script

To execute directly a bash script simply

docker run \
    --mount type=bind,source="$(pwd)",target=/home \
    dsblab/registrationtools:v0.1 /bin/bash -c "source activate registration: cd home; <bash_script.sh>"

and a python script

docker run \
    --mount type=bind,source="$(pwd)",target=/home \
    dsblab/registrationtools:v0.1 /bin/bash -c "source activate registration; cd home; python <python_script.py>"

⁠Jupyter lab

If you want to work interactively with a jupyter notebook.

docker run -it \
    -p 8888:8888 \
    --mount type=bind,source="$(pwd)",target=/home \
    dsblab/registrationtools:v0.1 /bin/bash -c "jupyter lab --notebook-dir=/home --ip='*' --port=8888 --allow-root"

You can then view the Jupyter Notebook by opening http://localhost:8888⁠ in your browser, or http://:8888 if you are using a Docker.