Your Docker image will be run with the commands listed at the bottom of the Methods page.
What to include with your submission:
- A ready-to-use Docker image with your method, renamed to
- A text file explaining:
- The command to execute in your container that will start the processing.
- Any special needs (GPU, specific Docker options, etc.)
- The expected runtime on CPU or GPU.
- The output of your method (either an 8-label or 3-label segmentation).
- A brief description of the method (1-2 pages)
Send your submission by email to: email@example.com. If your Docker image is too large to be sent by email, use a large file transfer service (e.g. WeTransfer, up to 2 GB). We can also send you a voucher for our large file transfer service (up to 500 GB) in advance.
Methods submitted for evaluation in this challenge will only be used by the organizers for this challenge.
Saving a Docker image
In case your image does not have the tag
mrbrains18/[TEAM-NAME] yet, then please use
docker tag to create a new tag.
docker save -o [TEAM-NAME].tar mrbrains18/[TEAM-NAME] to save your ready-to-use image as a tar archive. Optionally, compress the tar archive.
In case of errors, try reducing the number of layers in the image.
Your Docker image will be tested using the commands listed at the bottom of the Methods page. Specifically, we will run your container in the background (using
docker run -d) and then use
docker exec [YOUR-COMMAND] to start your method.
Please test this in advance on your own system.
Include a brief description of your method as a PDF document of 1-2 pages. This document should focus on the method itself and does not need an introduction, results, or discussion section.
Optionally, you can submit your method description to be published in the LNCS post-proceedings, please see additional details below.
Optionally, for participants who would like that, the description of the method can be extended to an 8-12 page manuscript after the challenge. This manuscript will be published in the LNCS post-proceedings of the joint BrainLes workshop (a collaboration between BrainLes, BraTS, ISLES, and the MRBrainS18 Challenge).
Submit your brief description, before the deadline, via the CMT submission system and select the MRBrainS18 track.
After the conference, the description of your method should be extended to a 8-12 page manuscript, including the evaluation and results of your method. The results will be kept secret until the conference, but then be announced and published on this website. Your manuscript will undergo (light) peer-review for correctness and completeness.