Created attachment 125
Minimal example which makes Skeletonize fails.

When applying the Skeletonize plugin to more or less vertical structures of size 2x2, it finds no skeleton.  Such an image would be for example (each slice being the same):

0 0 0 0 0 0
0 0 0 0 0 0
0 0 1 1 0 0
0 0 1 1 0 0
0 0 0 0 0 0
0 0 0 0 0 0

Working on it.

ignacio

Note:

A yet more minimal example suggesting that something may be wrong explicitly with the thinning from north or south:

0 0 0 0
0 0 0 0
0 1 1 0
0 0 0 0  -> fails

0 0 0 0
0 1 0 0
0 1 0 0
0 0 0 0  -> works

I wonder if that this is because the pixel removal needs to start somewhere and so this might a consequence of scanning the region in a particular way.

The IJ 2D skeleton has this characteristic too (try with a 2x2 region) you end up with a horizontal bar at the top, while there are potentially 4 different skeletons which are all correct.

This pattern ends in a point:
xx
xx
x

but this one ends in a 2x1 line:

xx
xx

and the point result above is different from "ultimate points" result.

All thinning methods start scanning from one side, which means they will give you a different result if you rotate your image for example. However this bug is something else, it should never remove a single voxel. I'll try to work on it in the next few weeks.

> However this bug is something else, it should never remove a single voxel.

Well, of course it should remove some voxels so that each xy slice contains at most a single white voxel.

You misunderstood me, what I meant is the algorithm should never remove a voxel if it is the only one that is still not deleted during the scanning. I will check if a condition is missing.