Regarding mismatching volume-label pairs in one subset

[anwai98]

Hi team,

Thanks for the CartoCell work.

I am not sure if this is the correct repo for asking the question, I'll go ahead anyways.

I spotted something: in the dataset hosted here, I see that in the low-resolution_MDCK-Hypoxia subset, 141/280 volumes have mismatching instance segmentation labels. Are we aware of this? If yes, is there a fix to it which you can point me to?

Thanks in advance!