Data Format and Coding for Free Viewpoint Video

In this paper we will discuss data representation formats and coding aspects of a new 3D functionality called Free Viewpoint Video (FVV). Similar to Computer Graphics (CG) or VRML applications, FVV allows for navigating with a virtual camera in 3D scenes. However, in contrast to graphical scenes in CG and VRML, FVV is based on natural scenes consisting of 3D video objects that provide a higher level of realism and allow a 3D reproduction of real events. Against this background the paper will review different approaches on FVV and different 3D representation formats. It will discuss the acquisition of required representation parameters and related rendering aspects. Furthermore it will present results of coding experiments for two particular 3D representation formats.