Virtual 3D camera system for training and media production
“Filmaking is a real craft, a miracle of collaboration”
Nowadays, most of the video content in films, video games, ads, is a combination of real and virtual elements. Therefore, previsualization is an important step in training and media production since it allows the producer or the camera operator to have an idea of the final composite. This can be achieved through a camera (virtual camera) that is tracked by sensors within a finite space (a.k.a. motion capture studio). In short, these systems also deal with a concept named mixed reality, where virtual and real objects coexist in the same space and interact between each other.
Currently, virtual camera operators and visual effects artists are constrained to few degrees of freedom and control over the captured content. As a result, the workflow is often compromised. Ideally, we would like to be able to extract key properties (e.g. color, placement) of any point in space. With this type of information everything on set and in post-production becomes more flexible. For instance, we no longer need the traditional chroma keying (green screen), it is easier to integrate visual effects, and we are able to refocus or change the view perspective after shooting the scene.
My mission is to improve each node of this complex system in order to meet the needs of artists. This includes developing a virtual 3D camera, and novel methods for multi-sensor calibration, rectification and multimodal data fusion.
Filipe has Bachelor’s and Master’s degrees in Electrical Engineering – Electronics and Telecommunications from Polytechnic Institute of Leiria (Portugal). During his Master’s studies, Filipe performed part of his thesis at the University of Erlangen-Nuremberg (FAU), Germany.
Department of Signal Processing
Tampere University of Technology
P. O. Box 553, FIN-33101 Tampere, Finland