Motion vector prediction

Patent metadata

Field	Value
Publication number	US-10602180-B2
Application number	US-201816003269-A
Country	US
Kind code	B2
Filing date	Jun 8, 2018
Priority date	Jun 13, 2017
Publication date	Mar 24, 2020
Grant date	Mar 24, 2020

Abstract

A video coder may determine a motion vector of a non-adjacent block of a current picture of the video data. The non-adjacent block is non-adjacent to a current block of the current picture. Furthermore, the video coder determines, based on the motion vector of the non-adjacent block, a motion vector predictor (MVP) for the current block. The video coder may determine a motion vector of the current block. The video coder may also determine a predictive block based on the motion vector of the current block.

First claim

What is claimed is:

1. A method of decoding video data, the method comprising:
storing, by a video decoder, a plurality of non-adjacent motion vector predictor (MVP) candidates in a first-in, first-out (FIFO) buffer;
determining, by the video decoder, a motion vector of a non-adjacent block of a current picture of the video data, the non-adjacent block being non-adjacent to a current block of the current picture, wherein the plurality of non-adjacent MVP candidates include a non-adjacent MVP candidate specifying the motion vector of the non-adjacent block;
determining, by the video decoder, based on the motion vector of the non-adjacent block, a MVP for the current block;
determining, by the video decoder, based on the MVP for the current block, a motion vector of the current block;
determining, by the video decoder, a predictive block based on the motion vector of the current block;
reconstructing, by the video decoder, based on the predictive block, sample values of the current picture; and
updating, by the video decoder, the FIFO buffer to remove an earlier-added, non-adjacent MVP candidate from the FIFO buffer and to add a new MVP candidate to the FIFO buffer.

2. The method of claim 1 , wherein determining the MVP for the current block comprises:
generating, by the video decoder, based in part on the motion vector of the non-adjacent block, a list of MVP candidates, wherein the list of MVP candidates includes an MVP candidate that specifies the motion vector of the non-adjacent block; and
determining, by the video decoder, from among the MVP candidates in the list of MVP candidates, the MVP for the current block.

3. The method of claim 2 , wherein the non-adjacent block is a first non-adjacent block, the method further comprising:
ordering, by the video decoder, a plurality of non-adjacent spatial MVP (NA-SMVP) candidates in the list,
wherein, for each respective NA-SMVP candidate of the plurality of NA-SMVP candidates, the respective NA-SMVP candidate corresponds to a respective non-adjacent block of a plurality of non-adjacent blocks, the respective NA-SMVP specifying a motion vector of the respective non-adjacent block,
wherein the plurality of non-adjacent blocks includes the first non-adjacent block, and
wherein the plurality of NA-SMVP candidates are ordered in the list according to distance of corresponding non-adjacent blocks from the current block.

4. The method of claim 2 , wherein the non-adjacent block is a first non-adjacent block, the method further comprising:
determining, by the video decoder, a first plurality of non-adjacent spatial MVP (NA-SMVP) candidates,
wherein, for each respective NA-SMVP candidate of the first plurality of NA-SMVP candidates, the respective NA-SMVP candidate corresponds to a respective non-adjacent block of a plurality of non-adjacent blocks, the respective NA-SMVP specifying the motion vector of the respective non-adjacent block,
wherein the first plurality of non-adjacent blocks includes the first non-adjacent block; and

ordering, by the video decoder, a second plurality of non-adjacent spatial MVP (NA-SMVP) candidates in the list, the second plurality of NA-SMVP candidates comprising non-duplicative NA-SMVP candidates in the first plurality of NA-SMVP candidates,
wherein the second plurality of NA-SMVP candidates are ordered in the list according to a frequency with which motion vectors specified by the NA-SMVP candidates in the second plurality of NA-SMVP candidates are specified by NA-SMVP candidates in the first plurality of NA-SMVP candidates.

5. The method of claim 2 , further comprising applying, by the video decoder, a pruning process to the list, wherein the pruning process is adaptive to a block size of the current block.

6. The method of claim 1 , wherein the MVP for the current block is in affine mode.

7. The method of claim 1 , wherein determining the motion vector of the current block comprises:
determining, by the video decoder, the motion vector of the current block such that the motion vector of the current block specifies a motion vector of the MVP for the current block.

8. The method of claim 1 , wherein determining the motion vector of the current block comprises:
determining, by the video decoder, the motion vector of the current block such that the motion vector of the current block is equal to a motion vector of motion information of the MVP for the current block plus a motion vector difference (MVD).

9. The method of claim 1 , wherein the non-adjacent block is a block in a template that defines a fixed pattern of blocks relative to a position of the current block.

10. The method of claim 1 , further comprising:
determining, by the video decoder, a set of coding unit (CU)-level frame-rate up conversion (FRUC) motion vector candidates;
selecting, by the video decoder, a CU-level FRUC motion vector candidate from the set of CU-level FRUC motion vector candidates;
determining, by the video decoder, a CU-level motion vector at least in part by performing a local search starting from a selected CU-level FRUC motion vector candidate; and
refining, by the video decoder, the CU-level motion vector at a sub-CU level with a set of FRUC sub-CU level motion vector candidates,
wherein at least one of the set of CU-level FRUC motion vector candidates and the set of FRUC sub-CU level motion vector candidates includes a non-adjacent spatial motion vector predictor (NA-SMVP) that specifies the motion vector of the non-adjacent block.

11. The method of claim 1 , further comprising:
prior to determining the motion vector of the non-adjacent block and based on a distance between the non-adjacent block and the current block being greater than a threshold distance, modifying a position of the non-adjacent block.

12. The method of claim 11 , wherein modifying the position of the non-adjacent block comprises rounding a position of the non-adjacent block to a position on a first grid of blocks in the current picture coarser than a second grid of blocks in the current picture.

13. The method of claim 11 , wherein modifying the position of the non-adjacent block comprises clipping a position of the non-adjacent block to the distance threshold.

14. The method of claim 1 , further comprising:
prior to determining the motion vector of the non-adjacent block and based on a distance between the non-adjacent block and a current coding tree block being greater than a threshold distance, modifying a position of the non-adjacent block, the current coding tree block containing the current block.

15. A method of encoding video data, the method comprising:
storing, by a video encoder, a plurality of non-adjacent motion vector predictor (MVP) candidates in a first-in, first-out (FIFO) buffer;
determining, by the video encoder, a motion vector of a non-adjacent block of a current picture of the video data, the non-adjacent block being non-adjacent to a current block of the current picture, wherein the plurality of non-adjacent MVP candidates include a non-adjacent MVP candidate specifying the motion vector of the non-adjacent block;
determining, by the video encoder, based on the motion vector of the non-adjacent block, a MVP for the current block;
determining, by the video encoder, a motion vector of the current block, wherein the motion vector is equal to a motion vector of the MVP for the current block or is equal to the motion vector of the MVP for the current block plus a motion vector difference (MVD) signaled in a bitstream;
determining, by the video encoder, a predictive block based on the motion vector of the current block;
generatin

Assignees

• Qualcomm Inc

Inventors

• Chen Yi-Wen
• Karczewicz Marta
• Chien Wei-Jung
• Sun Yu-Chen
• Zhang Li
• Lee Sungwon
• Li Xiang
• Chuang Hsiao-Chiang
• Chen Jianle
• Seregin Vadim

Classifications

• H04N19/44

  Decoders specially adapted therefor, e.g. video decoders which are asymmetric with respect to the encoder · CPC title

• H04N19/52 Primary

  by predictive encoding · CPC title

• H04N19/70

  characterised by syntax aspects related to video coding, e.g. related to compression standards · CPC title

• H04N19/593

  involving spatial prediction techniques · CPC title

• H04N19/176

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