/2023_challenge

BraTS 2023 Inpainting Challenge (Local Synthesis) Repository for Participants. Includes baseline model and infill mask generation script.

Primary LanguageJupyter NotebookMIT LicenseMIT

ATTENTION: The BraTS inpainting challenge 2023 is finished. We want to thank everybody who contributed to this event. Please follow us here, to stay updated regarding our BraTS inpainting challenges.

BraTS 2023 Inpainting Challenge (Local Synthesis)

Figure: Challenge Thumbnail

This repository is meant as tutorial for challenge participants.

If you have not visited the BraTS 2023 Inpainting Website yet, you should do so. Also considere reading the challenge manuscript for more context as the GitHub tutorial are rather technical in nature.

This repository is divided into three subtopics with a separate README file each.

  • Dataset: dataset/README.md
    Gives an overview of the challenge training dataset. Also includes the algorithm we used to generate the dataset.
  • Baseline Model: baseline/README.md Explains the creation of the baseline model. Also includes performance evaluations. You might want to use this chapter as starting point for your own (better) model!
  • Evaluation: evaluation/README.md Shows how out evaluation script works on the Synapse server using out baseline model as example. During the validation phase you will need to upload predictions to the Synapse server.
  • Submission: submission/README.md Guides you tough the synapse submission process using our baseline model as example.* This is relevant for the final submission where you will upload your whole model to Synapse.*