Automatic evaluation

This is short description with instruction notes how to create image to be upload to grand-challenge.org.
For newbies, please see Get started with Docker.
First you need to install Docker.io and run everything as super user sudo -i.

Build and export

For all operation with docker images you need to be the right user, or simply run sudo -i and continue with following commands.

Before you start, you need to adjust the ration of provided landmarks in Dockerfile using --min_landmarks parameter in Entry point.

Build

Copy the required data to the local directory and execute from actual location inside bm_ANHIR folder

docker build -t anhir -f Dockerfile .

Run and Test

Run one of following sample registration experiments:

simulate the ideal registration, assuming having all landmarks
bash python benchmark/bm_template.py \ -t ~/Medical-data/dataset_ANHIR/images/dataset_medium.csv \ -d ~/Medical-temp/dataset_ANHIR/images \ -o ~/Medical-temp/experiments_anhir/ \ -cfg sample_config.yaml python bm_experiments/bm_comp_perform.py -o ~/Medical-temp/experiments_anhir/BmTemplate # remove all registered images rm ~/Medical-temp/experiments_anhir/BmTemplate/*/*.jpg \ ~/Medical-temp/experiments_anhir/BmTemplate/*/*.png
run bUnwarpJ in ImageJ registration on the real data
bash python bm_experiments/bm_bUnwarpJ.py \ -t ~/Medical-data/dataset_ANHIR/images/dataset_medium.csv \ -d ~/Medical-temp/dataset_ANHIR/images \ -o ~/Medical-temp/experiments_anhir/ \ --run_comp_benchmark \ -Fiji ~/Applications/Fiji.app/ImageJ-linux64 \ -cfg ./configs/ImageJ_bUnwarpJ_histol-1k.txt # remove all registered images rm ~/Medical-temp/experiments_anhir/BmUnwarpJ/*/*.jpg \ ~/Medical-temp/experiments_anhir/BmUnwarpJ/*/*.png

Running the docker image with mapped folders

mkdir submission output

and upload the sample submission to submission and run the image

docker run --rm -it \
    --memory=4g \
    -v $(pwd)/submission/:/input/ \
    -v $(pwd)/output/:/output/ \
    anhir

Export

Export the created image to be uploaded to the evaluation system.

# full size image
docker save anhir > anhir.tar
# compressed image
docker save anhir | gzip -c > anhir.tar.gz

Browsing and cleaning

Browsing
To see your locally build images use:

docker image ls

Cleaning
In case you fail with some builds, you may need to clean your local storage.

docker system prune

or Docker - How to cleanup (unused) resources

docker images | grep "none"
docker rmi $(docker images | grep "none" | awk '/ / { print $3 }')