Submissions for the Preliminary Phase are open!

Hello everyone,

I would like to begin by expressing my sincere appreciation for your interest in the upcoming challenge.

I am excited to announce that the Preliminary Phase has officially started, and I invite you all to start submitting your algorithms. During this phase, you will have the opportunity to test your algorithms on a subset of the test volumes, specifically 3 out of the total 50. It is of utmost importance to ensure that your algorithm, including the docker container if applicable, functions correctly and to gain initial insights into its performance. For detailed information on the dates for the opening and closing of the submission phases, please visit the Important Dates page.

Additionally, I have recently updated our example repository by introducing a simplified version of the PosPadUNet3D. The purpose of this code is to provide you with a customizable foundation while omitting intricate details. To provide a baseline for comparison, I have also submitted the example code to the Challenge Leaderboard, to let you have some baseline scores.

If you have any questions or concerns, please feel free to ask.

Best regards, Luca Lumetti

Dear Organizer, Could you please tell me the format path of the input and the output.

For example, if the CBCT data I choose to test is P100.mha and P101.mha.
Input Dir: /input/images/cbct/P100.mha; /input/images/cbct/P101.mha
Out Dir: /output/images/inferior-alveolar-canal/P100.mha; /output/images/inferior-alveolar-canal/P101.mha

As I have submitted the docker image following the format above, but I got the information when submitting that the algorithm failed on one or more cases. Whether the format path of the input and the output above is actually wrong ?

I should have already responded to you by email, but I would also like to answer this question also here for further participants who read it. The input/output settings you described look right. To have more detailed information on how the example code we provided works (which is based on the evalutils library) I suggest you to dig into their codebase, specifically the SegmentationAlgorithm class which we extend by implementing the predict method.