| 摘要 |
A multiview video decoding apparatus receives and decodes an encoded stream obtained as a result of encoding residual information, 2D images for N views, and depth images for N views, the residual information being the error between synthetic-view images generated using the 2D images for N views and the depth images for N views, and 2D images for (M−N) views at the view synthesis positions of the synthetic-view images. A view synthesizing apparatus generates the synthetic-view images by using the 2D images and depth images for N views decoded by the multiview video decoding apparatus. A residual information compensating apparatus adds the residual information into the generated synthetic-view images. The apparatus may be applied to a system that conducts view synthesis, for example.
|
