Gregory Lee

Thanks for raising the issue @zoccoler. Indeed, 3D regions with an extent of 1 pixel along one of the dimensions seems like a good thing to add test cases for...

> I am not very familiar with ImageJ but a I couldn't find a medial surface thinning implementation. Maybe someone more familiar with it could confirm this? The [Skeletonize3D plugin](https://imagej.net/plugins/skeletonize3d)...

> In more details, for the model estimation within the RANSAC, it seems to take src and dst arguments that has size of (N, 2) and each row is (col,...

Per https://github.com/scikit-image/scikit-image/issues/1234#issuecomment-516079885, we need to decide on `rescale_to_*` vs. `convert_to_*`. I prefer `rescale_to` because it is clearer that it rescales the range of the input and is not just a...

We will need to hold off on merging this PR until after we branch 1.0

> There's a fourth column, which is a modified version of this PR: > > P, M, N when there is a 3-spatial-dimensions image, K, M, N when it's a...

Hi @iliailmer, Thank you for the contribution! I have not yet done a detailed review of the publication, but it has a substantial number of citations and the implementation looks...

It seems okay to me, if you can also update the docstring to reference the alternative form used when `alpha` is provided (assuming that variant is not already discussed in...

may need to revert 4b63878 unless we update to SciPy >=1.6.0

Thanks @crisluengo! I am in favor of making this API change for `skimage2`. As you know, we have already had multiple people confused by the current state of things. On...