Picture prediction method and apparatus, and codec

Patent metadata

Field	Value
Publication number	US-11528503-B2
Application number	US-202016915678-A
Country	US
Kind code	B2
Filing date	Jun 29, 2020
Priority date	Dec 31, 2017
Publication date	Dec 13, 2022
Grant date	Dec 13, 2022

Abstract

A picture prediction method is provided, which includes: obtaining initial motion information of a current picture block; determining, based on a matching cost criterion, that positions of a pair of reference blocks are a position of a target forward reference block of the current picture block and a position of a target backward reference block of the current picture block, where positions of each pair of reference blocks include a position of a forward reference block and a position of a backward reference block; and for the positions of each pair of reference blocks, a first position offset and a second position offset are in a mirror relationship; and obtaining a predicted value of a pixel value of the current picture block based on a pixel value of the target forward reference block and a pixel value of the target backward reference block.

First claim

What is claimed is:

1. A picture prediction method, comprising:
obtaining initial motion information of a current picture block;
when an early termination condition is not met, determining, from positions of N forward reference blocks and N backward reference blocks, based on a matching cost criterion, positions of a target forward reference block and a target backward reference block of the current picture block, wherein the positions of N forward and N backward reference blocks are based on the initial motion information, the N forward reference blocks comprising an initial forward reference block, the N backward reference blocks comprising an initial backward reference block, and N is an integer greater than 1,
wherein for positions of n-th forward and n-th backward reference blocks of the N forward and N backward reference blocks, a first position offset and a second position offset are in a mirror relationship, the first position offset representing an offset of the position of the n-th forward reference block relative to a position of the initial forward reference block, and the second position offset representing an offset of the position of the n-th backward reference block relative to a position of the initial backward reference block; wherein n is an integer and 1≤n≤N; and

obtaining a predicted value of a pixel value of the current picture block based on a pixel value of the target forward reference block and a pixel value of the target backward reference block;
wherein the positions of the N forward reference blocks comprise a position of the initial forward reference block and positions of (N−1) candidate forward reference blocks, and an offset of a position of each candidate forward reference block relative to the position of the initial forward reference block is an integer pixel distance; and
the positions of the N backward reference blocks comprise a position of the initial backward reference block and positions of (N−1) candidate backward reference blocks, and an offset of a position of each candidate backward reference block relative to the position of the initial backward reference block is an integer pixel distance.

2. The method according to claim 1 , wherein the position of the target forward reference block is associated with (delta0x′, delta0y′) representing an offset of the position of the target forward reference block relative to the position of the initial forward reference block,
the position of the target backward reference block is associated with (delta1x′, delta1y′) representing an offset of the position of the target backward reference block relative to the position of the initial backward reference block, and
wherein delta0x′=−delta1x′, and delta0y′=−delta1y′.

3. The method according to claim 1 , wherein the positions of the N forward reference blocks comprise a position of the initial forward reference block and positions of (N−1) candidate forward reference blocks; or
the positions of the N backward reference blocks comprise a position of the initial backward reference block and positions of (N−1) candidate backward reference blocks,
for positions of n-th candidate forwad and n-th candidate backward reference blocks of the (N−1) candidate forward and (N−1) candidate backward reference blocks, a first position offset and a second position offset are in a mirror relationship, the first position offset representing an offset of the position of the n-th candidate forward reference block relative to a position of the initial forward reference block, and the second position offset representing an offset of the position of the n-th candidate backward reference block relative to a position of the initial backward reference block n is an integer and 0<n≤N−1.

4. The method according to claim 1 , wherein a direction of the first position offset is opposite to a direction of the second position offset, and an amplitude of the first position offset is the same as an amplitude of the second position offset.

5. The method according to claim 1 , wherein the first position offset is represented by (delta0x, delta0y) and the second position offset is represented by (delta1x, delta1y), wherein delta0x=−deltalx, and delta0y=−delta1y.

6. The method according to claim 1 , further comprising:
obtaining updated motion information of the current picture block, wherein the updated motion information comprises an updated forward motion vector and an updated backward motion vector,
wherein the updated forward motion vector points to the position of the target forward reference block, and the updated backward motion vector points to the position of the target backward reference block, or
wherein the updated forward motion vector indicates an offset of the position of the target forward reference block relative to the position of the current picture block, and the updated backward motion vector indicates an offset of the position of the target backward reference block relative to the position of the current picture block.

7. The method according to claim 1 , wherein the pixel value of the target forward reference block is determined based on the position of the target forward reference block, and the pixel value of the target backward reference block is determined based on the position of the target backward reference block.

8. The method according to claim 1 , wherein the initial motion information comprises a first motion vector and a first reference picture index corresponding to a first list (L0), and a second motion vector and a second reference picture index corresponding to a second list (L1); and
wherein the determining positions of N forward and N backward reference blocks comprises:
determining, based on the first motion vector and the position of the current picture block, the position of the initial forward reference block of the current picture block in a forward reference picture corresponding to the first reference picture index, using the position of the initial forward reference block as a first search start point, and determining the positions of the (N−1) candidate forward reference blocks in the forward reference picture, wherein the positions of the N forward reference blocks comprise the position of the initial forward reference block and the positions of the (N−1) candidate forward reference blocks; and
determining, based on the second motion vector and the position of the current picture block, the position of the initial backward reference block of the current picture block in a backward reference picture corresponding to the second reference picture index, using the position of the initial backward reference block as a second search start point, and determining the positions of the (N−1) candidate backward reference blocks in the backward reference picture, wherein the positions of the N backward reference blocks comprise the position of the initial backward reference block and the positions of the (N−1) candidate backward reference blocks.

9. The method according to claim 1 , wherein the determining, from the positions of the N forward and N backward reference blocks based on a matching cost criterion, positions of a target forward reference block and a target backward reference block of the current picture block, comprises:
determining, from the positions of the N forward and N backward reference blocks, the positions of the target forward and target backward reference blocks of the current picture block, wherein the target forward and target backward reference blocks have a minimum matching error among N matching errors of the N forward and N backward reference blocks; or
determining, from the positions of the N forward and N backward reference blocks, the positions of the target forward and ta

Assignees

• Huawei Tech Co Ltd

Inventors

• Ma Xiang
• Yang Haitao
• Chen Huanbang
• Gao Shan

Classifications

• H04N19/577 Primary

  Motion compensation with bidirectional frame interpolation, i.e. using B-pictures · CPC title

• H04N19/56 Primary

  Motion estimation with initialisation of the vector search, e.g. estimating a good candidate to initiate a search · CPC title

• H04N19/176

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

• H04N19/52 Primary

  by predictive encoding · CPC title

• H04N19/105

  Selection of the reference unit for prediction within a chosen coding or prediction mode, e.g. adaptive choice of position and number of pixels used for prediction · CPC title