Encoding and decoding methods and apparatuses for enhancement layer

Patent metadata

Field	Value
Publication number	US-12457334-B2
Application number	US-202318206794-A
Country	US
Kind code	B2
Filing date	Jun 7, 2023
Priority date	Dec 8, 2020
Publication date	Oct 28, 2025
Grant date	Oct 28, 2025

Abstract

An encoding method includes: an encoder obtains a reconstructed block of a base layer of a to-be-encoded picture block; calculates a difference between corresponding pixels in the to-be-encoded picture block and the reconstructed block of the base layer to obtain a residual block of an enhancement layer of the to-be-encoded picture block; determines a transform-block partitioning manner of the residual block of the enhancement layer; and performs transformation on the residual block of the enhancement layer based on the transform-block partitioning manner. This application can simplify a processing procedure for the encoder and improve encoding efficiency of the encoder. In addition, compression efficiency of the residual block can be improved more effectively by using an adaptive TU partitioning method.

First claim

We claim:

1. An encoding method, comprising:
obtaining a reconstructed block of a base layer of a to-be-encoded picture block;
calculating a difference between corresponding pixels in the to-be-encoded picture block and the reconstructed block of the base layer to obtain a residual block of an enhancement layer of the to-be-encoded picture block, wherein a resolution of an enhancement layer picture block is not lower than a resolution of a base layer picture block, or an encoding quality of the enhancement layer picture block is not lower than an encoding quality of the base layer picture block;
determining a transform-block partitioning manner of the residual block of the enhancement layer, including:
performing iterative tree structure partitioning on a first largest transform unit (LTU) of the residual block of the enhancement layer to obtain transform units (TUs) of a plurality of partitioning depths, wherein a maximum partitioning depth of the plurality of partitioning depths is equal to Dmax, and Dmax is a positive integer; and
determining a partitioning manner of a first TU based on a loss estimate of the first TU and a sum of loss estimates of a plurality of second TUs comprised in the first TU, wherein a partitioning depth of the first TU is i, a depth of the second TU is i+1, and 0≤i≤Dmax−1, wherein the transform-block partitioning manner of the residual block of the enhancement layer is different from a transform-block partitioning manner of a residual block of the base layer; and

performing transformation on the residual block of the enhancement layer based on the transform-block partitioning manner to obtain a bitstream of the residual block of the enhancement layer.

2. The method according to claim 1 , wherein the determining the partitioning manner of the first TU based on the loss estimate of the first TU and the sum of the loss estimates of the plurality of second TUs comprised in the first TU comprises:
performing transformation and quantization on a TU to obtain a quantized coefficient of the TU, wherein the TU is the first TU or one of the plurality of second TUs;
precoding the quantized coefficient of the TU to obtain a code word length of the TU;
performing dequantization and inverse transformation on the quantized coefficient of the TU to obtain a reconstructed block of the TU;
calculating a sum of squared differences (SSD) of the TU and the reconstructed block of the TU to obtain a distortion value of the TU;
obtaining a loss estimate of the TU based on the code word length of the TU and the distortion value of the TU;
determining a smaller one of the loss estimate of the first TU and a sum of the loss estimates of the plurality of second TUs; and
determining the partitioning manner corresponding to the smaller one as the partitioning manner of the first TU.

3. The method according to claim 1 , wherein a size of the first LTU is the same as a size of the reconstructed block of the base layer, or a size of the first LTU is the same as a size of a reconstructed block obtained by performing upsampling on an original reconstructed block of the base layer.

4. The method according to claim 1 , wherein the to-be-encoded picture block is a largest coding unit (LCU) in a full-frame picture, the to-be-encoded picture block is a full-frame picture, or the to-be-encoded picture block is a region of interest (ROI) in a full-frame picture.

5. The method according to claim 1 , wherein the base layer and the enhancement layer are obtained based on resolution or encoding quality scalability.

6. The method according to claim 1 , wherein when the base layer and the enhancement layer are obtained based on resolution scalability, before the calculating the difference between the corresponding pixels in the to-be-encoded picture block and the reconstructed block of the base layer to obtain the residual block of the enhancement layer of the to-be-encoded picture block, and
wherein the method further comprises:
performing downsampling on an original to-be-encoded picture block to obtain the to-be-encoded picture block with a first resolution; and
performing upsampling on an original reconstructed block of the base layer to obtain the reconstructed block of the base layer with the first resolution.

7. The method according to claim 1 , wherein the obtaining the reconstructed block of the base layer of the to-be-encoded picture block comprises:
calculating a difference between corresponding pixels in the to-be-encoded picture block and a prediction block of the to-be-encoded picture block to obtain a residual block of the to-be-encoded picture block;
performing transformation and quantization on the residual block of the to-be-encoded picture block to obtain a quantized coefficient of the to-be-encoded picture block;
performing dequantization and inverse transformation on the quantized coefficient of the to-be-encoded picture block to obtain a reconstructed residual block of the to-be-encoded picture block; and
performing summation on corresponding pixels in the prediction block of the to-be-encoded picture block and the reconstructed residual block of the to-be-encoded picture block to obtain the reconstructed block of the base layer.

8. A decoding method, comprising:
obtaining a bitstream of a residual block of an enhancement layer of a to-be-decoded picture block, wherein the bitstream of the residual block of the enhancement layer is encoded by determining a transform-block partitioning manner of the residual block of the enhancement layer, including performing iterative tree structure partitioning on a first largest transform unit (LTU) of the residual block of the enhancement layer to obtain transform units (TUs) of a plurality of partitioning depths, wherein a maximum partitioning depth of the plurality of partitioning depths is equal to Dmax, and Dmax is a positive integer, and determining a partitioning manner of a first TU based on a loss estimate of the first TU and a sum of loss estimates of a plurality of second TUs comprised in the first TU, wherein a partitioning depth of the first TU is i, a depth of the second TU is i+1, and 0≤i≤Dmax−1;
performing entropy decoding, dequantization, and inverse transformation on the bitstream of the residual block of the enhancement layer to obtain a reconstructed residual block of the enhancement layer;
obtaining a reconstructed block of a base layer of the to-be-decoded picture block, wherein a resolution of an enhancement layer picture block is not lower than a resolution of a base layer picture block, or an encoding quality of the enhancement layer picture block is not lower than an encoding quality of the base layer picture block; and
performing summation on corresponding pixels in the reconstructed residual block of the enhancement layer and the reconstructed block of the base layer to obtain a reconstructed block of the enhancement layer, wherein the transform-block partitioning manner of the residual block of the enhancement layer is determined to be different from a transform-block partitioning manner of a residual block of the base layer.

9. The method according to claim 8 , wherein the to-be-decoded picture block is a largest coding unit (LCU) in a full-frame picture, the to-be-decoded picture block is a full-frame picture, or the to-be-decoded picture block is a region of interest (ROI) in the full-frame picture.

10. The method according to claim 8 , wherein the base layer and the enhancement layer are obtained through layering in a quality scalability manner or a spatial scalability manner.

11. The method according to claim 8 , wherein when the base layer and the enhancement layer are obtained through layering in a sp

Assignees

• Huawei Tech Co Ltd

Inventors

• Song Yi
• Di Peiyun
• Chen Shaolin
• Zhang Yixuan

Classifications

• H04N19/96

  Tree coding, e.g. quad-tree coding · CPC title

• H04N19/30 Primary

  using hierarchical techniques, e.g. scalability (H04N19/63 takes precedence) · CPC title

• H04N19/176

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

• H04N19/13

  Adaptive entropy coding, e.g. adaptive variable length coding [AVLC] or context adaptive binary arithmetic coding [CABAC] · CPC title

• H04N19/147

  according to rate distortion criteria (rate-distortion as a criterion for motion estimation H04N19/567) · CPC title