When a sampling point is located near some part of a mesh, it is utterly useless to compute distances to triangles that are “on the other side” of the mesh. One might think that the concept of “closeness” to some triangles of the mesh is inseparable from the distance we intend to compute, but it turns out that there is another set of approaches utilizing a “friend” of programmers who want to accelerate a process handling arbitrarily large amounts of data – a binary tree
Martin Cavarga 