Classification submission question

Hi!

I've an idle question about the classification task. How exactly are the scores calculated? As we only submit the case label, I assumed the score is the accuracy and the AUC is being disregarded. However, the scores are far too granular for the 30 validation cases. How exactly are the scores for the classification calculated?

Cheers, Jan

Scores are based on the code in our GitHub's repository https://github.com/PerceptionComputingLab/TDSC-ABUS2023/tree/main/Metrics. While the validation phase has its metrics, the test phase might differ. We'll provide detailed submission requirements soon.