Name
Abstract | ||
|---|---|---|
In this work, we present a model-aided coder that incorporates 3-D shape and motion information to improve the performance of a video coding sys- tem. The 3-D shape model of a rigid body is esti- mated from the video sequence directly. Using the shape model, the 3-D motion of the rigid body and the background global motion are also estimated for video coding. Experimental results show that bit-rate savings of 30%-40% are achieved at equal PSNR compared to the H.26L test model TML- 8.0. This corresponds to 2-3 dB improvements in PSNR when encoding at the same bit-rate. ered. 3-D shape and motion of each rigid body is estimated, and each frame is synthesized from the reconstructed previous frame using the 3-D information. The residual in regions that the model does not predict well is additionally coded. In this work, we consider video sequences con- taining one rigid object of primary interest. No assumption is made for other parts of the scene. The intrinsic camera parameters are assumed to be known and remain constant through the sequence. For video coding, the 3-D shape of the object is estimated and used in the model-aided coding framework. The 3-D motion of the object as well as the global motion in the background is esti- mated to facilitate motion-compensated prediction. Experimental results are compared with the H.26L test model TML-8.0 coder (5). 2 Model-Aided Coding |
Year | Venue | Keywords |
|---|---|---|
2002 | vision modeling and visualization | rigid body |
Field | DocType | Citations |
Computer vision,Block-matching algorithm,Quarter-pixel motion,Computer graphics (images),Computer science,Motion compensation,Rigid body,Coding (social sciences),Artificial intelligence,Motion estimation,D-Shape,Encoding (memory) | Conference | 0 |
PageRank | References | Authors |
0.34 | 9 | 3 |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Chuo-Ling Chang | 1 | 247 | 18.15 |
| Peter Eisert | 2 | 461 | 55.59 |
| Bernd Girod | 3 | 8988 | 1062.96 |