Diapositive 13 sur 19
Commentaires:
When the tool moves, we could use the previous technique detecting collisions at discrete tools position.
But in such a case, we can miss facets that collided with the tool between two consecutive time steps. Taking these facets into account or not leads to very different deformed states for the organ and to wrong force feedback information.
If the movement is fast or if the object is small, we can even totally miss a collision.
To avoid this, we consider the collision between the organ and the volume described by the tool during the last simulation step.
As in endoscopy the tool passes through a fixed point (the aperture in the patient’s skin), this volume is like a piece of camembert.
OpenGL provides a second type of camera which is useful to approximate this kind of volume, the perspective camera.
