Dear Participants,

We hope your algorithm development is progressing well, and we look forward to receive your submissions. Here's a quick update on the challenge timeline, along with a compilation of resources to aid your algorithm development.

Challenge Timeline

  • December 6, 2023: KickOff and release of revised ThoraxCBCT data.
  • January 31, 2024: Docker submission deadline.
  • February 20, 2024: Virtual Workshop presenting results and announcing winners.
  • Post-workshop: Journal submission preparation.

Submission Guidelines

  • We accept containerised algorithms runnable exclusively via Docker.
  • For deep learning solutions, both training and inference components must be included.
  • The container must not include additional datasets for training, as we retrain your solution on a hidden dataset.
  • Incorporate logging for tracking progress in data loading and training, and easier debugging.
  • Infrastructure details: We provide a 20GB NVIDIA RTX 4000 SFF; ensure your solution adheres to these specifications.
  • Expect the hidden dataset to share the same image format as the revised ThoraxCBCT dataset.
  • The file structure for the hidden data will mirror that of ThoraxCBCT, with only the prefix changing from ‘ThoraxCBCT_’ to ‘OncoReg_’.
  • Base your data loading on the '_dataset.json' available with our previous datasets.
  • Provide clear instructions on building and running your container in your submission.
  • Regarding a planned journal submission, please provide us with a structured description of your method if you would like to be featured.

Visit our GitHub repository, where we provide an exemplary OncoReg submission using VoxelMorph++. Find more information on containerisation there, and use the example as a template for your submission.

In an effort to avoid unnecessary resource usage, we've opted not to conduct sanity checks on submissions requiring multiple hours of training. Thus, we kindly request you to provide a minimal version of your solution capable of delivering results within one hour. Submit this downscaled version prior to your final submission for a sanity check to ensure that we will be able to rank your method. We will not provide any qualitative results regarding registration performance before your final submission.

We will be accepting submissions until January 31, 2024 11:59 pm under this upload link. Ensure you upload a compressed version of your Docker repository, accompanied by comprehensive documentation (think of it as a well-written GitHub readme).


Results and insights will be presented at our virtual workshop scheduled on February 20, 2024, from 2:30 – 5 pm CET. Successful participants will have the opportunity to present their work, and winners will be awarded price money. Further information will follow in due course.

We look forward to your innovative submissions!

Best regards,
Wiebke & the OncoReg Team