Fast Mesh-based Coding of Depth Map Sequences for Efficient 3D-Video Reproduction using OpenGL

S. Grewatsch and E. Müller (Germany)


3D-video, depth maps, mesh coding, OpenGL


During the reproduction of 3D-video based on monoscopic video and corresponding depth information, two virtual views are generated in real-time for binocular perception. This has been done by depth-image-based rendering tech niques. In order to make this step more efficiently, the au thors applied the capabilities of current graphics hardware to the view rendering. Therefore the platform independent OpenGL library was used. To accomplish this approach the depth maps were converted into 3D-meshes. However, the mesh generation was computationally expensive. The pur pose of this research was to discuss different methods of obtaining the meshes more efficiently. For this reason cur rent results of human-factors experiments were taken into account. Finally, a strategy for fast mesh-based lossy com pression of depth map sequences was developed.

Important Links:

Go Back