Rules

Submission

• Submit your solution to validation phases with a limit until they close (to validate your submissions technically)
  • Submit to Validation CTA
  • Submit to Validation MRA
• 🏆 Submit your final best solution ✨ for private test data during test phases (only one submission per phase)
  • Submit to Final Test CTA
  • Submit to Final Test MRA

Please use the validation submission portals to make sure your Docker container works for the submission platform: reads from the correct input interface and saves to the correct output interface. Validation phases also help you get an idea of the evaluation format. The validation images are from the training set, selected purely based on size to trigger any resource-related errors. You can submit to validation phases with a daily limit until they close. But final test phases only allow for one submission.

Account and Team

• Each participant can join at most one team.
• Normally, each participant/team can only submit one algorithm to each track.
  • If you or your team have sufficiently distinct algorithms, you may submit more than one algorithm to a modality track, subject to approval from us.
  • Please reach out to us if want to split into sub-teams or submit multiple algorithms.
• We will exclude duplicate submissions from final leaderboards.
• We will exclude anonymous submissions unless we can verify your profile.

Inference Environment Restrictions

The inference runtime environment on Grand-challenge cloud platform has the folowing setup:

• 1x NVIDIA T4 GPU with 16 GiB VRAM
• Main memory of at most 31 GiB DRAM ⚠️
• You can specify the GPU and Memory options in your Algorithm's "Update Settings". More info here
• Maximum 15 minutes per image for now (can be extended to 20 minutes if this is a bottleneck for most participants)
• More info on the cloud platform can be found here.
• Your Docker container size must be under 10GB ⚠️

Please take note of the above inference restrictions when designing your algorithm.

Usage of Training Data

Participants may use any other public datasets and private in-house data, or modify the supplied TopBrain 2025 training data, provided that they disclose and mention any additional or modified training datasets in their description of the submitted algorithm (please see "Publication policy" below).

Members of the Organizing Institutes

Members of the organizers' direct research groups can participate and their results can be included in the publications and the leaderboard. However, they are not eligible for awards. This policy also applies to teams that include such members.

Award and Result Announcement Policy

Top 3 teams (excluding the organizer teams) from each of the two tracks will be publicly named and given a certificate along with a booklet "Anatomy of the Brain Illustrated Pocket Anatomy" 🧠 and a Swiss wooden cow toy 🐄 as souverniors at the in-person challenge event.

All participants/teams are invited to prepare a 4-minute presentation/video for the challenge session to present and discuss their methods. If you are joining the event, please fill up the contact form soon. See below.

After the public announcement, a detailed analysis of the submitted results will be available upon request.

🏆 📰 Publication Policy

The challenge results will be summarized and published in a journal manuscript. All participants with a (reasonable) submission are invited to contribute to our challenge publication!

In order for us to include you in our paper, please:

Each submission can have maximum three co-authorships for the challenge paper. Additional authors from the top submissions can be included upon request with justification according to the ICMJE authorship guidelines.

Participating teams may submit their results separately without any publication embargo.