In the current implementation, we assign "1" at the top of C2 (below the superior part of C1) and "2" at the midpoint between levels 1 and 2. When only part of C1 is present in the image, this creates an issue as shown in the attached image.
To solve this:
- We can project the superior and inferior voxels of C1 as levels 1 and 2, respectively (possibly after cropping only the part of C1 that is anterior to the canal anterior-line).
- After adding brainstem segmentation as suggested in issue #48, we can put level 1 only if the brainstem is present.
- This solution will be more effective after improving the C1 ground truth segmentation, as it is currently not reliable enough for setting level 2.
In the current implementation, we assign "1" at the top of C2 (below the superior part of C1) and "2" at the midpoint between levels 1 and 2. When only part of C1 is present in the image, this creates an issue as shown in the attached image.
To solve this: