Submission Guidelines¶

Participants must submit their solution as a Docker container through the Grand Challenge platform (https://cd24.grand-challenge.org/). An example of the creation of the docker container is available on github: https://github.com/TheodoreAouad/dc-sfr-2024-example.

The Docker container is run on every input image concurrently. Please go and have a look at the example repository to understand how to create the Docker container and how the input data are read. Please ensure that your output has type "uint8".

Evaluation¶

The evaluation script is open source and available on github: https://github.com/TheodoreAouad/dc-sfr-2024-evaluation. The final score is computed by aggregating four metrics. Our purpose is first to evaluate the segmentation results to ensure the anatomical coherence of the results, and then to measure the total volume predicted as it has a high clinical relevance. We measure both these metrics on the whole prediction, as well as on the eloquent areas of the brain, which are the most important for the clinical outcome. The final score is a weighted sum of these four metrics.

For more details, please refer to the open source repository https://github.com/TheodoreAouad/dc-sfr-2024-evaluation.