Question from a beginner
3D collision detection is a little different because you need to choose how accurate you want to be.
If you have say a person one way is to use a bounding box, which is basicly a box fit right around the person and anything that fits in the box is considered a collision.
Though this collision detection is fast it is not accurate. Say the person has thier arms stretched out and the object colliding was underneath them, you will still get a collision because of the box around the person will catch it. (This is easier to show , than write)
So you have a choice, you can use boinding boxes for each part of a person or a more accurate, polytope.
This is basicly a very simplified version of the objects 3D mesh. You can than use this for collisions with more accuracy.
Gary