Collision and containment detection between biomechanically based eye muscle volumes.

Collision and containment detection between three-dimensional objects is a common requirement in simulation systems. However, few solutions exist when exclusively working with deformable bodies. In our ophthalmologic diagnostic software system, the extraocular eye muscles are represented by surface models, which have been reconstructed from magnetic resonance images. Those models are projected onto the muscle paths calculated by the system's biomechanical model. Due to this projection collisions occur. For their detection, three approaches have been implemented, which we present in this paper: one based on image-space techniques using OpenGL, one based on the Bullet physics library and one using an optimized space-array data structure together with software rendering. Finally, an outlook on a possible response to the detected collisions is given.