Transform Block Size Determination in Video Coding

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

In existing video coding and decoding technologies, the size of a transform matrix is not correlated with the size of a prediction block, leading to abrupt transforms across prediction block boundaries, which weakens the transform effect and reduces coding efficiency.

Innovation Solution

A method is introduced where the size of a transform block is determined by considering the dividing manner of an image block, and identification information that identifies the corresponding dividing layer number of the transform block is written into a code stream, ensuring that the transform block does not cross the boundary of the prediction block, thereby improving coding efficiency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If the size of transform block is not correlated with prediction block size, then transform operation can be performed independently, but transform block crosses prediction block boundary causing weakened transform effect and reduced coding efficiency

Engineering Contradiction:
Improvecoding efficiencyVSAvoidtransform effect
Core Design Contradiction:
ProductivityVSManufacturing precision

Solution Approach 1:

The transform block is segmented to align with prediction block boundaries. The coding unit is divided into multiple prediction blocks, and transform blocks are correspondingly segmented to match these boundaries, preventing cross-boundary transforms and improving both transform effect and coding efficiency

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Different transform block sizes are applied to different regions based on local prediction block characteristics. Each prediction block region receives transform processing tailored to its specific size and characteristics, optimizing the transform effect locally while maintaining overall coding efficiency

Inventive Principle:
Principle #3Local quality

2Device complexity

If transform block size is fixed independent of prediction block, then processing is simpler, but transform correlation removal is insufficient

Engineering Contradiction:
Improveprocessing complexityVSAvoidcorrelation removal
Core Design Contradiction:
Device complexityVSLoss of information

Solution Approach 1:

The transform block size is made dynamic and adaptive to the prediction block size. Rather than using a fixed transform block size, the system dynamically adjusts transform block dimensions to match the corresponding prediction block, enabling effective correlation removal while maintaining manageable processing complexity through systematic adaptation

Inventive Principle:
Principle #15Dynamics

Data Source

PatentEP3157256B1Method for coding an image comprising determining transform block size
Publication Date: 2019.10.09 HUAWEI TECH CO LTD
  • EP3157256B1 patent drawingFigure 1a~1c
  • EP3157256B1 patent drawingFigure 1d~2b
  • EP3157256B1 patent drawingFigure 2c~3a

AI summary

The present invention provides a method for obtaining a size of a transform block, where the method includes: obtaining a dividing manner of an image block; and obtaining size information of a post-dividing transform block according to the dividing manner of the image block and a width and a height of a pre-dividing transform block. By using the method for obtaining a size of a transform block provided in this embodiment of the present invention, a size of a transform block may be coded according to a dividing manner of an image block or a height and a width of a pre-dividing transform block, thereby effectively improving coding efficiency.