Projective camera model in biomedical navigation applications

Surgical interventions pose very high demands on the spatial sense of surgeons. Especially in complex surgeries, navigation support is an important factor for the success of an operation. With the introduction of computers in surgical navigation, the field of computer assisted surgery was born. In difference to former mechanical constructions, these systems are highly adaptive and enable the development of versatile medical applications. This thesis examines different aspects of integrating a camera into surgical 3D navigation systems. Depending on the context, a camera can perform a self-localization or has to be tracked. In either case, this information is used to build intuitive visualization interfaces, where the operation planning and the reality are fused in a spatially correct manner. The accuracy and reliability of such augmented reality navigation applications is a major concern in this thesis. However, besides the high level incorporation of existing algorithms, this thesis also investigates the projective transformation at its core. It is shown that the position and behavior of a controllable tilting mirror can be described compactly with the projective camera model.