For the HECKTOR 2025 challenge, evaluation will be conducted exclusively through Docker container submissions on the Grand Challenge platform. Docker Submission Overview
The task submission is divided into 1+2 phases:
- Sanity Check Phase: Consists of 3 images to ensure participants are familiar with the Grand Challenge platform and that their dockers run without errors. All teams must make their submission to this phase and will receive feedback on any errors.
- Validation Phase: Consists of approximately 50 images. All teams will submit up to 2 working dockers from the sanity check to this phase. Only the top 15 teams, as ranked by the evaluation metrics displayed on the public validation leaderboard, with valid submissions will proceed to Phase 3.
- Testing Phase: Consists of approximately 400 images. The teams will choose 1 of their 2 dockers from the validation phase to be submitted to the testing phase. The official ranking of the teams will be based solely on the testing phase results.
The participants will not receive detailed feedback during the testing phase except for error notifications.
Task-Specific Submission Requirements¶
- Task 1: Docker containers should produce segmentation outputs as a single label mask per patient (1 for the predicted GTVp, 2 for GTVn, and 0 for the background) in
.mha
format. The resolution of this mask should be the same as the original CT resolution. Participants should ensure correct pixel spacing and origin with respect to the original reference frame. The mha files should be named[PatientID].mha
, matching the patient names, e.g.,CHUB-001.mha
. - Task 2: Docker containers should produce results as a JSON file with the prediction output of the model as a float. The output should be anti-concordant with the RFS in days (i.e., the model should output a predicted risk score).
- Task 3: Docker containers should produce results as a JSON file containing the output of the model (boolean). The output should be True for HPV positive and False for HPV negative.
🚨 Important Submission Details 🚨¶
All participants must adhere to the following constraints during submission:
🕒 Inference Time Limits per Task¶
- Task 1: Maximum 10 minutes
- Task 2: Maximum 15 minutes
- Task 3: Maximum 10 minutes
📦 Container Requirements¶
- Maximum container size: 10 GB
💻 Compute Resources (per submission)¶
- Instance Type: ml.g4dn.2xlarge
- GPU: 1× NVIDIA T4 (16 GiB VRAM)
- CPU: 8 vCPUs
- RAM: 32 GiB
- /tmp storage: 225 GB NVMe SSD
Please make sure your containers run efficiently within these limits.
📦 Docker Submission Resources Now Available¶
To support participants in preparing their submissions, we have released detailed documentation and example code for Docker-based inference on our official HECKTOR 2025 GitHub repository.
The repository is organized to simplify your development workflow:
main branch
: Contains baseline implementations and inference pipelines for all three challenge tasks. Use these as starting points or references for your own models.docker-template branch
: Includes a comprehensive guide for building, testing, and packaging your Docker containers. Each task has a dedicated example with the expected input/output format and helper scripts to validate your setup locally before submission.
We strongly encourage all teams to review the repository carefully and build their containers using the provided structure to avoid compatibility issues during evaluation.
Please check the repository regularly for any updates or clarifications.