Re: Howto model decoupled hierarchies?

It not the vertex / point storing this data. It's at contour-level.
Means for each contour we calculate the convex hull and kD-tree.
- A convex hull is some kind of simplified contour (it resembles some
kind of deformed circle) and also have lots of vertices (the vertices
of the contour).
- A kD-Tree is an "ordered" set of the vertices that allows fast
discovery of the points next to any arbitrary other point. The kD-tree
is actually a 2D-tree

Would the following data structure meet your app's basic requirements:

point1
x
10
y
25
z
54
closestPoints
->point7
->point3
->point56
->point43
...
convexHull
->point7
->point56
->point56
->point3
...

where x, y, z are raw data
and closestNodes and convexHull are derived data.
Do you need to related a point to more than one convexHull at the same
time?

.

Relevant Pages

  • Re: Howto model decoupled hierarchies?
    ... calculated data might be the convex hull of the contour (closed ... section of your mesh into some product at the highest level, ... hull, KD-Tree, simple array - doesn't matter). ... I the convex hull, kd-tree and other temponaries are just the first ...
    (comp.object)
  • Re: Howto model decoupled hierarchies?
    ... Means for each contour we calculate the convex hull and kD-tree. ... kind of deformed circle) and also have lots of vertices (the vertices ...
    (comp.object)
  • Re: Howto model decoupled hierarchies?
    ... so each point stores 2 types of calculated data. ... Means for each contour we calculate the convex hull and kD-tree. ...
    (comp.object)