Monday, April 28, 2014

[ejmsyyvx] Voronoi averaging

Calculate the Voronoi diagram of a set of points.  Define a new set of points as the centroid of each of the Voronoi partitions.  Repeat.

Unclear of what to do about the regions of infinite size.  Maybe those points don't move.

Closely spaced points will probably get moved away from each other.  What will it converge to?  A different center than centroid, perhaps incenter? (I suspect it is easy to calculate the largest circle that can be inscribed within a convex polygon but am not sure.)  Can the steps be reversed?

No comments :