Method for sub-PU motion information inheritance in 3D video coding

#The present invention is a national stage of PCT/CN2015/071606, filed Jan. 27, 2015, which claims priority to PCT Patent Application, Serial No. PCT/CN2014/071576, filed on Jan. 27, 2014, entitled “Methods for Motion Parameter Hole Filling”. The PCT Patent Application is hereby incorporated by reference in its entirety.

#The present invention relates to three-dimensional video coding. In particular, the present invention relates to depth coding using motion information inherited from texture pictures in three-dimensional video coding.

#Three-dimensional (3D) television has been a technology trend in recent years that intends to bring viewers sensational viewing experience. Various technologies have been developed to enable 3D viewing. The multi-view video is a key technology for 3DTV application among others. The traditional video is a two-dimensional (2D) medium that only provides viewers a single view of a scene from the perspective of the camera. However, the multi-view video is capable of offering arbitrary viewpoints of dynamic scenes and provides viewers the sensation of realism. 3D video formats may also include depth maps associated with corresponding texture pictures. The depth maps also have to be coded to rendering three-dimensional view or multi-view.

#FIG. 1 illustrates an example of motion hole filling process according to the existing three-dimensional video coding based on High Efficiency Video Coding (3D-HEVC).

#FIG. 2A illustrates an example of motion hole filling process to use the motion parameters of a co-located texture block associated with the center sub-PU corresponding to the sub-PU at a lower-right position of the center point of the PU to determine default motion parameters according to an embodiment of the present invention.

#FIG. 2B illustrates an example of the default motion parameter derivation based on motion parameters of the co-located texture block covering the center sample of the sub-PU at a lower-right position of the center point of the PU, where the location of the center sample of the sub-PU corresponds to a lower right sample of the center point of the sub-PU.

#In the conventional 3D-HEVC, the Motion Parameter Inheritance (MPI) hole filling process for a motion hole relies on the last available MPI before the motion hole. Therefore, the motion parameters of a motion hole are dependent on the sub-PU processing order, which is not friendly to parallel processing. To overcome this issue, embodiments of the present invention use default motion parameters for motion hole filling so that motion prediction for each sub-PU can be done in parallel. Furthermore, pruning of motion candidates does not need to wait for all sub-PUs to determine inherited motion parameters. Accordingly, for each sub-PU within the current PU, if its corresponding reference block does not contain available motion parameters, default motion parameters are used for the sub-PU. The motion information associated with a PU or sub-PU includes one or more of the motion vector(s), reference picture index(es), reference picture list(s), etc. The motion information associated with each PU or sub-PU is referred as motion parameters in this disclosure.

#The default motion parameters may be determined according to the motion parameters of the co-located texture reference block associated with a selected depth sub-PU. Furthermore, the co-located texture reference block may correspond to a reference block covering a center sample of the selected depth sub-PU. The selected depth sub-PU may correspond to any sub-PU in the depth PU. However, it is preferred to select a depth sub-PU in the center of the PU. For example, the center sub-PU may correspond to the sub-PU at a lower-right position (220) of the center point (210) of the PU (230) as shown in FIG. 2A. Nevertheless, other center depth sub-PU may also be selected. For example, the center sub-PU of the PU may correspond to the sub-block located in a lower-left, upper-right or upper-left position of the center point (210). In other words, the center sub-PU corresponds to the sub-PU in the respective quadrant and adjacent to the center point (210) of the current PU. While a prediction unit (PU) is used as an example of block structure, other block structure such as coding unit (CU) or macroblock (MB) may also be used.

#One example of determining the default motion parameters is disclosed as follows. If the co-located texture block covering the center of a depth PU has available motion parameters, the available motion parameters are used as the default motion parameters. For example, if nPSW and nPSH are the width and height of the PU, and nSubPsW and nSubPsH are the width and height of the sub-PU, the center pixel (xc, yc) can be determined according to: xc=(nPSW/nSubPsW/2)*nSubPsW+nSubPsW/2, and  (1) yc=(nPSH/nSubPsH/2)*nSubPsH+nSubPsH/2).  (2)