Improving child particles interpolation

Currently the interpolation between strands is made linearly in the 3D space, this is usually good enough for most cases but, when there is the need to have thin meshes or the number of hair isn't high enough, there could be unpleasant glitches that need a lot of work to be fixed or, in most cases they can't be fixed

I'd propose to use the geodesic distance (or some of its variations, like Dijkstra, diffusion distance and so) in mesh space instead of the euclidean distance in object space. The important part is to have the distances of the guide strands based on the topology of the mesh. Luckily there are already functions present in the CGAL library that can overcome to this problem

Also: the interpolation itself can be improved if the interpolation was done with the geometric transformation based on the chapter 5 of this paper

• A Practical Model for Hair Mutual Interactions