| 摘要 |
A depth map in a three dimensional [3D] video signal is processed. From the 3D video signal a first depth map (Z1) is derived. A second depth map (Z2) is generated by a multi dimensional filter (22) that causes the second depth map to have spilling artifacts whereas the first depth map in corresponding locations has less or no such artifacts. A depth difference is determined between the first depth map and the second depth map a positive depth difference indicating a depth in the second depth map being closer to a viewer. A final third depth map is generated by combining first depth map values and second depth map values according to a combining function in dependence of the depth difference. The combining function gives preference to the first values where the depth difference is positive. |
