rib morphometry: Extract metrics from the centerline

[msrepo]

[msrepo]

rib-enclosed area

[msrepo]

There is a data quality issues with this dataset. It looks like the segmentation was performed by simple thresholding of CT hounsfield unit (HU).

For ground truth generation, we used existing morphology-based algorithms and manually refined its results. Then, considering the sparsity of ribs in 3D volumes, we thresholded and sampled sparse voxels from the input and designed a point cloud-based baseline method for rib segmentation.

from https://zenodo.org/record/5336592

https://arxiv.org/pdf/2109.09521.pdf

[msrepo]

Yes, there seems to have been manual cleaning done to avoid scapula and other nearby regions being wrongly annotated as rib (which is good), but at the expense of some sections of the rib being wiped out, may be because that section of the rib consisted of less dense (non-cortical bone).

let's not get into data cleaning. Use the centerline from the original dataset to obtain the required metrics.