Good evening everyone. I’m using data from https://www.cancerimagingarchive.net and I’m having trouble using the data. I notice that I have a single file for the segmentations in dicom format and when I import it into Python I have many more segmentation slices than the CT slices. How can this situation be managed and why does it happen? Thanks a lot
As long as the segmentation is correct in physical space, everything is good. If you need the segmentation geometry (origin, spacing, axis directions, extents) to match any particular volume - for example because you want to process it with some software that ignores physical space - then you can resample the segmentation by a few clicks (Segmentations model, Export, Labelmap, choose a reference volume).
Yes I converted them from dicom into nrrd and it works.
Thank you.