Multi View Imaging (MVI) techniques are finding extensive usage in future multimedia, surveillance, inspection and video conferencing applications. Particularly, the multi-billion dollar entertainment/film industry is spending huge amounts in R & D related to immersive multimedia (e.g. Immersive TV). However, the biggest hurdle to-date has been the parallel processing of multiple video streams acquired in real time. This limits the benefits of MVI based systems for most applications.
The project proposes to design and develop a prototype system aimed as a platform for Multi View Imaging (MVI) applications such as Immersive Television and Video Conferencing. The project will produce a complete embedded hardware system, capable of generating real time panoramic mosaics, using video input from multiple cameras. The MVI platform will also provide basic user interaction capabilities to customize the generated panoramic video, which can assist in solving another problem in immersive TV content generation, namely the Free Viewpoint TV (FTV in Immersive TV terminology).
The prototype is being developed on a reconfigurable architecture, hence it can then be used to prototype other deliverable products useful for the local market and export. These applications encompass Internet (Panoramic video conferencing, Tele-monitoring of an object), Media (User-defined view for sporting events, Time freeze-videos), Security (Advanced surveillance system allowing to view and compress video from multiple cameras in a single view, Real time panoramic views for armed personal vehicles). The reconfigurable MVI platform will also provide a test-bed for the academia for rapid development of video processing based research applications.