Point Cloud Clustering for Interlinked Objects

Hi all, I'm about to do an experiment which involves capturing a 3D point cloud of interlinking chains, stitched and reconstructed from all-around 3D scans. Here is a physical sample of how the 5 interlinking chains will look like; I have yet to obtain the point cloud.

I'm trying to find a way / algorithm which will allow me to segment the chains, e.g. chain 1, chain 2, ... chain 5. This will allow me to just present the point cloud of a single chain on its own when I call for it.

An example of a clustering algorithm would be the k-means but I believe this technique will not do well in separating the chains at the linkage interface.

Does anyone have any idea / able to direct me to anything useful?

Thanks a lot!

[prashik shende]

I am also looking for same solution did you get somthing?