Problems with Skeletonize for Ellipsoid Factor

I am having trouble with skeletonizing prior to the ellipsoid factor analysis.

First, I import an image sequence (BMP files) of a rat tibia metaphyseal trabecular VOI which I have created using CTan from Bruker.

I then manually threshold (image>adjust>threshold) the image using the ‘red’ setting before I click apply and try to match it as well as I can to the original trabecular geometries. I then select ‘black and white’ plus ‘dark background’ before clicking apply. This binarizes my image nicely.

Next, I select (plugins>skeleton>skeletonize 2D/3D) and it runs fine.

The resulting skeletons are largely good quality, however circles appear every so often, almost always around the outer edges of the cross section (the endosteal perimeter of the bone marrow cavity). When viewed in 3D, these circles appear as spheres or ‘lumps’ in my otherwise nice looking skeletonized dataset.

What is causing these? Do I call them skeletonizing artifacts? How do I eliminate them? Do I need to eliminate them? Are there any other pre-processing steps which I have missed out and need to include?

Best regards,


please do a 3d binary fill befor you skeleton.
the solid would result in skeleton, and the holes would be bubble.
this is a demo in 2D show why:

and if you want to cut some small branch, you can can try:
ImagePy Network Simplified

Ok I’ll try it out. Thanks very much for your help.