I used TotalSegmentator on a CT image, and it successfully segmented all relevant structures. However, I noticed that it does not include a specific label for air that I need.
I need to determine the HU range—minimum and maximum voxel values—that correspond to air in the original CT image. Could anyone please explain how I can extract or visualize this information using 3D Slicer?
Air is usually -1000 (Hounsfield scale - Wikipedia) and you can segment with simple thresholding. But if you are inside the body be aware there’s a lot of partial volume effects (a voxel has air and tissue) so it can be challenging to get an exact volume.
Thank you for your help! I have another question — do you happen to know how I can assign labels to different segments (obtained from the Total Segmentator tool) and save them together in a single .mhd file? Is that possible?
I’d really appreciate any guidance on how to do this properly.
I’m not sure what you mean. It is possible to export selected segments (the visible ones) to a labelmap and then save that in mhd format. If that’s not what you mean please describe what you are trying to accomplish.
Yeah, I have tried doing that. But the generated file doesn’t save the original name I had for each segmentation. It gives the names like the ones for the label:
I followed the steps and exported the segmentation in the format I wanted. However, when I load the file back into 3D Slicer, the segment names aren’t preserved—it behaves the same way as before, without showing the original segment names.
I saw that when I go to the data and select segmentsss-label, I am able to see how I want the name next to the label. It’s when I save the file and reopen it separately that the name disappears, so I am confused why this happens.
A segmentation node which contains segments (as produced by say TotalSegmentator) by default can be saved to a .seg.nrrd file format which saves the entire segmentation node (with all the segments) and that can be reloaded back into Slicer with the segment names you specified.
Yes, I’m aware of that, but the issue is that it doesn’t show the label values for each segment—such as the liver, which I specifically need. What I am trying to do is to create a map where I can define which label value corresponds to which anatomical structure (e.g., label value 3 = liver, etc.).