Multi-view coding and view synthesis for 3DTV

In this paper we analyze the current developments in 3DTV technologies concerning multi-view video coding and free-viewpoint rendering. Multi-View Coding (MVC) can be performed in several ways that influence the performance and the complexity of the system. We underline the connection between coding and free-viewpoint rendering techniques. Since the latest research shows that a better multi-view coding can be reached by using interview prediction based on view synthesis, we present our view synthesis algorithm and compare it with the MPEG standardization software. The first quality measurements, indicating the difference between our view synthesis and the MPEG algorithm, are provided and discussed.