Resurrecting an ancient thread here, hope that’s kosher.
Anyone worked on turning those papers into gh components at all?
My hack-ish approach to contribute is: voxelize and the use tri-remesh with a bounding box or mesh sphere (does it make a difference?) as the starting geometry and the voxel mesh as the target. The result is definitely not a proper concave hull as the vertices will not be coincident with original points, nor will all the points be enclosed. But it works to get a nice bean-shaped thing that generalizes a point cloud shape.
I made it for turning tree canopy LiDAR data into volumes for CFD.
