Inter prediction method and apparatus
AssigneeHuawei Technologies Co., Ltd.
InventorsHuanbang Chen, Haitao Yang, Lian Zhang
Life of this patent
Estimating the term…
Abstract
An inter prediction method, including: when a prediction mode of a to-be-processed picture block is an affine motion model-based advanced motion vector prediction mode, obtaining control point motion vectors of the to-be-processed picture block, where the control point motion vectors meet preset first motion vector resolution and/or first motion vector bit depth; then, deriving a motion vector of each motion compensation unit in the to-be-processed picture block based on the control point motion vectors; and obtaining a reconstructed block of the to-be-processed picture block based on the motion vector of each motion compensation unit.
Claims25
A video codec device, comprising:
- a memory containing instructions;
- and
- a processor in communication with the memory and, upon execution of the instructions, is configured to perform the operations of:
- determining that a prediction mode of a to-be-processed picture block is not an affine motion model-based merge mode;
- obtaining control point motion vectors of the to-be-processed picture block, wherein the control point motion vectors have a preset first motion vector resolution and/or a first motion vector bit depth, wherein the obtaining control point motion vectors of the to-be-processed picture block comprises:
- obtaining control point motion vector differences (CPMVDs) and the control point motion vector predictors (CPMVPs);
- and
- based on a resolution of the CPMVDs being not equal to the first motion vector resolution, obtaining the control point motion vectors of the to-be-processed picture block based on the CPMVDs and a first shift value;
- or
- based on a resolution of the CPMVPs being not equal to the first motion vector resolution, obtaining the control point motion vectors of the to-be-processed picture block based on the CPMVPs and a second shift value;
- deriving a motion vector of each motion compensation unit in the to-be-processed picture block based on the control point motion vectors of the to-be-processed picture block;
- and
- obtaining a reconstructed block of the to-be-processed picture block based on the motion vector of each motion compensation unit in the to-be-processed picture block.
A video codec device, comprising:
- a memory containing instructions;
- and
- a processor in communication with the memory and, upon execution of the instructions, is configured to perform the operations of:
- determining that a prediction mode of a to-be-processed picture block is an affine motion model-based merge prediction mode;
- obtaining control point motion vectors of the to-be-processed picture block;
- adjusting the control point motion vectors to meet a preset first motion vector resolution and/or a first motion vector bit depth, wherein the adjusting the control point motion vectors comprises:
- based on a resolution of the control point motion vectors being not equal to the first motion vector resolution, obtaining the adjusted control point motion vectors based on the control point motion vectors and a first shift value;
- deriving a motion vector of each motion compensation unit in the to-be-processed picture block based on adjusted control point motion vectors;
- and
- obtaining a reconstructed block of the to-be-processed picture block based on the motion vector of each motion compensation unit.
An inter prediction method, comprising:
- determining that a prediction mode of a to-be-processed picture block is not an affine motion model-based merge mode;
- obtaining control point motion vectors of the to-be-processed picture block, wherein the control point motion vectors have a preset first motion vector resolution and/or a first motion vector bit depth, wherein the obtaining control point motion vectors of the to-be-processed picture block comprises:
- obtaining control point motion vector differences (CPMVDs) and the control point motion vector predictors (CPMVPs);
- and
- based on a resolution of the CPMVDs being not equal to the first motion vector resolution, obtaining the control point motion vectors of the to-be-processed picture block based on the CPMVDs and a first shift value;
- or
- based on a resolution of the CPMVPs being not equal to the first motion vector resolution, obtaining the control point motion vectors of the to-be-processed picture block based on the CPMVPs and a second shift value;
- deriving a motion vector of each motion compensation unit in the to-be-processed picture block based on the control point motion vectors of the to-be-processed picture block;
- and
- obtaining a reconstructed block of the to-be-processed picture block based on the motion vector of each motion compensation unit in the to-be-processed picture block.
An inter prediction method, comprising:
- determining that a prediction mode of a to-be-processed picture block is an affine motion model-based merge prediction mode;
- obtaining control point motion vectors of the to-be-processed picture block;
- adjusting the control point motion vectors to meet a preset first motion vector resolution and/or a first motion vector bit depth, wherein the adjusting the control point motion vectors comprises:
- based on a resolution of the control point motion vectors being not equal to the first motion vector resolution, obtaining the adjusted control point motion vectors based on the control point motion vectors and a first shift value;
- deriving a motion vector of each motion compensation unit in the to-be-processed picture block based on adjusted control point motion vectors;
- and
- obtaining a reconstructed block of the to-be-processed picture block based on the motion vector of each motion compensation unit.
Dependencies are inferred from the claim text.