Sub-PU motion prediction for texture and depth coding

Patent metadata

Field	Value
Publication number	US-9948915-B2
Application number	US-201414339256-A
Country	US
Kind code	B2
Filing date	Jul 23, 2014
Priority date	Jul 24, 2013
Publication date	Apr 17, 2018
Grant date	Apr 17, 2018

Abstract

In accordance with one or more techniques of this disclosure, a video coder may divide a current prediction unit (PU) into a plurality of sub-PUs. Each of the sub-PUs may have a size smaller than a size of the PU. Furthermore, the current PU may be in a depth view of the multi-view video data. For each respective sub-PU from the plurality of sub-PUs, the video coder may identify a reference block for the respective sub-PU. The reference block may be co-located with the respective sub-PU in a texture view corresponding to the depth view. The video coder may use motion parameters of the identified reference block for the respective sub-PU to determine motion parameters for the respective sub-PU.

First claim

What is claimed is:

1. A method of decoding multi-view video data, the method comprising:
dividing, by a video decoder, a current prediction unit (PU) of a current coding unit (CU) of a current picture into a plurality of sub-PUs, each of the sub-PUs having a size smaller than a size of the current PU, the current PU being in a depth view of the multi-view video data;
deriving, by the video decoder, default motion parameters, for the current PU, from a texture block co-located with a center sub-PU of the current PU;
for each respective sub-PU of the plurality of sub-PUs:
identifying, by the video decoder, a reference block for the respective sub-PU, wherein the identified reference block for the respective sub-PU is co-located with the respective sub-PU and the identified reference block for the respective sub-PU is in a texture view corresponding to the depth view,
when motion parameters of the identified reference block for the respective sub-PU are available, using, by the video decoder, motion parameters of the identified reference block for the respective sub-PU to determine motion parameters for the respective sub-PU, wherein the motion parameters of the identified reference block comprise a motion vector,
when the motion parameters of the identified reference block for the respective sub-PU are not available, setting the motion parameters of the respective sub-PU to the default motion parameters, and
using, by the video decoder, the motion parameters for the respective sub-PU to determine a respective predictive block for the respective sub-PU;

determining, by the video decoder, a predictive block for the current PU by assembling the predictive blocks for the sub-PUs; and
reconstructing, by the video decoder, the current PU based at least in part on the predictive block for the current PU.

2. The method of claim 1 , wherein for each respective sub-PU of the plurality of sub-PUs, the motion parameters of the identified reference block for the respective sub-PU include a first motion vector, a second motion vector, a first reference index, and a second reference index, the first motion vector and the first reference index being for a first reference picture list, the second motion vector and the second reference index being for a second reference picture list.

3. The method of claim 1 , further comprising:
including, by the video decoder, a particular candidate in a merge candidate list of the current PU, wherein the particular candidate has the motion parameters of each of the sub-PUs;
obtaining, by the video decoder, from a bitstream, a syntax element that indicates a selected candidate in the merge candidate list; and
based on the selected candidate being the particular candidate, invoking, by the video decoder, motion compensation for each of the sub-PUs.

4. The method of claim 1 ,
wherein the default motion parameters comprise a first default motion vector, a second default motion vector, a first default reference index, and a second default reference index, the first default motion vector and the first default reference index being for a first reference picture list, the second default motion vector and the second default reference index being for a second reference picture list.

5. The method of claim 1 , wherein each sub-PU of the plurality of sub-PUs has a block size equal to 4×4, 8×8, or 16×16.

6. The method of claim 1 , wherein using the motion parameters of the identified reference block for the respective sub-PU to determine the motion parameters for the respective sub-PU comprises using, by the video decoder, the motion parameters of the identified reference block for the respective sub-PU as the motion parameters for the respective sub-PU.

7. The method of claim 1 , further comprising:
dividing the current CU of the current picture into a plurality of PUs, the plurality of PUs including the current PU.

8. The method of claim 1 , wherein the center sub-PU is the sub-PU with the following coordinates relative to a top-left sample of a prediction block of the current PU:
((( nPSW >>( u+ 1))−1)<< u , (( nPSH >>( u+ 1))−1<< u )

or with the following coordinates relative to a top-left sample of a prediction block of the current PU:
(( nPSW >>( u+ 1))<< u , (( nPSH >>( u+ 1))<< u )

where nPSW and nPSH are the width and height of the prediction block of the current PU, and where the sub-PU size is 2 U ×2 U .

9. A method of encoding multi-view video data, the method comprising:
dividing, by a video encoder, a current prediction unit (PU) of a current coding unit (CU) of a current picture into a plurality of sub-PUs, each of the sub-PUs having a size smaller than a size of the current PU, the current PU being in a depth view of the multi-view video data;
deriving, by the video encoder, default motion parameters, for the current PU, from a texture block co-located with a center sub-PU of the current PU;
for each respective sub-PU of the plurality of sub-PUs:
identifying, by the video encoder, a reference block for the respective sub-PU, wherein the identified reference block for the respective sub-PU is co-located with the respective sub-PU and the identified reference block for the respective sub-PU is in a texture view corresponding to the depth view,
when motion parameters of the identified reference block for the respective sub-PU are available, using, by the video encoder, motion parameters of the identified reference block for the respective sub-PU to determine motion parameters for the respective sub-PU, wherein the motion parameters of the identified reference block comprise a motion vector,
when the motion parameters of the identified reference block for the respective sub-PU are not available, setting the motion parameters of the respective sub-PU to the default motion parameters, and
using, by the video encoder, the motion parameters for the respective sub-PU to determine a respective predictive block for the respective sub-PU;

determining, by the video encoder, a predictive block for the current PU by assembling the predictive blocks for the sub-PUs; and
encoding, by the video encoder, the current PU based at least in part on the predictive block for the current PU.

10. The method of claim 9 , wherein for each respective sub-PU of the plurality of sub-PUs, the motion parameters of the identified reference block for the respective sub-PU include a first motion vector, a second motion vector, a first reference index, and a second reference index, the first motion vector and the first reference index being for a first reference picture list, the second motion vector and the second reference index being for a second reference picture list.

11. The method of claim 9 , further comprising:
including, by the video encoder, a particular candidate in a merge candidate list of the current PU, wherein the particular candidate has the motion parameters of each of the sub-PUs;
signaling, by the video encoder, in a bitstream, a syntax element that indicates a selected candidate in the merge candidate list; and
based on the selected candidate being the particular candidate, invoking, by the video encoder, motion compensation for each of the sub-PUs.

12. The method of claim 9 ,
wherein the default motion parameters comprise a first default motion vector, a second default motion vector, a first default reference index, and a second default reference index, the first default motion vector and the first default reference index being for a first reference picture list, the second default motion vector and the second default reference index being for a second reference picture list.

13. The method of

Assignees

• Qualcomm Inc

Inventors

• Chen Ying
• Zhang Li

Classifications

• H04N19/176

  the region being a block, e.g. a macroblock · CPC title

• H04N19/597 Primary

  specially adapted for multi-view video sequence encoding · CPC title

• H04N19/513

  Processing of motion vectors · CPC title

• H04N19/51

  Motion estimation or motion compensation · CPC title

• H04N13/0048 Primary

  Electricity · mapped topic