The data provided for this challenge is anonymized according to current regulations (removal of DICOM information as well as the facial part in the anatomical image data).
We received approval from the ethics committee of Charité – Universitätsmedizin Berlin under ID EA2/222/19 .
Datasets are made available to registered users under CC BY-NC-ND 4.0 license, meaning that data should only be shared between registered users, neither redistribution nor change or reuse is allowed. This license is only valid as long as regulations for data anonymization do not change.
The code for calculating the validation metrics and their statistics is developed using a python and/or MeVisLab-backend in the COMIC environment.
A specification of the underlying calculations will be provided via the challenge web page.
Teams decide themselves whether they want to provide links to code repositories together with the description (paper) of their approach.
Conflicts of interest
The challenge is sponsored by the German Federal Ministry for Education and Research ((Berlin Institute for the Foundations of Learning and Data (BIFOLD)) and the Fraunhofer Society.
The test labels will be available to registered participants according to the announced timeline described above.