CONVERSE: scalable representations for image compression and view synthesis
CONVERSE builds on the success of a large European project, VIRTUE (2000-2003), which developed the first immersive videoconferencing system prototype in Europe based entirely on image processing and synthesis techniques. CONVERSE takes forward VIRTUE work by exploring the role of models of the human figure for arbitrary-viewpoint rendering (ako free-viewpoint TV). CONVERSE is a collaboration with the Department of Electronic Engineering of the University of Bristol (Prof Nishan Canagarajah). The project explores also novel compression techniques for head-shoulders-torso sequences (e.g., newscasting, videoconferencing)within and beyond the MPEG-4 and X.26x frameworks, addressing scalability,3-D rendering, and multiple-view video compression and rendering.
Funder: EPSRC
Start Date: 17/01/2005
End Date: 16/01/2008
For further information about the CONVERSE: scalable representations for image compression and view synthesis project please contact Professor Emanuele Trucco.