Participation
This page collects the practical information participants need most: where to register, what to submit, and the main dates for the challenge workflow.
How to Participate
- Register on the Synapse challenge page and accept the Data Usage Agreement (EndoVis rules); every person on the team should sign.
- Download the filtered iMED dataset associated with your task and familiarize yourself with the data format and baseline implementations (links on the Getting Started page).
- Develop your method. You may use publicly available external data for pretraining or co-training, but private/non-public clinical data is not permitted.
- Validate by submitting to the public validation leaderboard (opens June 30).
- Submit your Docker container by the deadline. Task 1 containers must implement
get_pose(); Task 2 containers must implementnew_view(). - Submit a structured methods report by September 1.
Use this page for registration and challenge updates.
Downloading the data
You can download the data manually under the Files tab on Synapse. Alternatively, you can download programmatically:
import synapseutils
import synapseclient
email = "user@email.edu" # Your email address in the Synapse account
authToken = "123456789" # add your authentication
local_folder = "./sample_data" # Your local folder for the downloaded data
syn = synapseclient.Synapse()
syn.login(authToken=authToken)
files = synapseutils.syncFromSynapse(syn, 'syn74277461', path=local_folder)
Signed rules requirement
After registering on Synapse, each participant must email one signed copy of the EndoVis rules form (one signed form per person) to the primary organizer, Sierra Bonilla, at sierra.bonilla.21@ucl.ac.uk. Download the form here: EndoVis Rules PDF.
Participant timeline
- April 15, 2026 Registration opens and the dataset is released on Synapse.
- May 15, 2026 Docker Containers for subtasks will be released on Synapse.
- June 30, 2026 Public validation leaderboard opens.
- September 1, 2026 Structured methods report due.
- September 27, 2026 Challenge Day at MICCAI 2026 / EndoVis Strasbourg, France.
Checklist before submission
1. Register your team
Create a team entry on Synapse, accept the Data Usage Agreement, and make sure your contact email is correct.
2. Validate your Docker image
Test your container with the official loader and submit to the public validation leaderboard once it opens.
3. Document external data use
List all external datasets, pretrained checkpoints, and additional supervision sources in the method report.
4. Email your signed rules form
Each participant must email one signed copy of the EndoVis rules form to the primary organizer, Sierra Bonilla, at sierra.bonilla.21@ucl.ac.uk.
5. Submit your structured report
Submit a structured method report by September 1 with your approach, training/inference pipeline, and any external data or pretrained models used.
FAQ
Can we use external public datasets?
Yes. The current policy allows publicly available datasets and pretrained networks, provided they are clearly documented and legally accessible.
Can we use private hospital data?
No. Private clinical data and non-public institutional datasets are not permitted.
Will there be a public validation phase?
Yes. A public validation phase will provide official scripts, baseline code, and intermediate submissions before final evaluation.
Will code need to be open-sourced?
Open-sourcing is encouraged but not required. Teams must submit executable Docker containers to the organizers.
Contact
For questions about registration, submissions, or participation, contact Sierra Bonilla. Additional official support channels will be listed here once available.