The Digilut challenge, a collaborative project between Trustii.io and Foch Hospital and supported by the Health Data Hub and Bpifrance, aims to develop a medical decision support algorithm to diagnose graft rejection episodes in lung transplant patients by detecting pathological regions (A lesions) on transbronchial biopsies.
Install poetry:
brew install poetry
Clone the project:
git clone https://github.com/LouisStefanuto/digilut.git
Create a virtual environment and install deps:
cd digilut
poetry install --with dev,docs
For a more detailed install, refer to the install section of the documentation.
The datasets are available on the competition webpage. Note that you have to create a Trustii account first.
- 412 annotated slides (~200 GB): needed
- 1680 non-annotated slides (~3TB): not used in this project
For more info about the project, visit the open documentation.
You can also host it locally by running:
mkdocs serve
If you encounter any issues with the repo or have suggestions for improvements, feel free to open an issue or submit a pull request.