Merge Candidate Transform Inheritance for Video Block Coding

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing video coding technologies face inefficiencies in determining transform type inheritance for blocks in merge modes, particularly in inter and intra block copy merge modes, leading to suboptimal compression and decoding performance.

Innovation Solution

A method and apparatus for video encoding and decoding that involve processing a bitstream to determine whether transform type information is inherited by a current block based on predefined conditions, adding transform type information of a merge candidate and motion information into a merge candidate list when applicable, and encoding or reconstructing the block based on this list.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If transform type information is inherited by the current block in merge mode, then compression efficiency is improved, but the complexity of determining inheritance conditions increases

Engineering Contradiction:
Improvecompression efficiencyVSAvoidcomplexity of determining inheritance conditions
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent applies parameter changes by establishing specific conditions based on block size (width and height dimensions) to determine when transform type information should be inherited. By changing the parameters of block dimension thresholds, the system automatically selects appropriate inheritance scenarios without requiring complex decision logic, thus improving compression efficiency while maintaining manageable complexity.

Inventive Principle:
Principle #35Parameter changes

2Ease of operation

If transform type information is always inherited in merge mode, then decoding simplicity is improved, but compression performance deteriorates

Engineering Contradiction:
Improvedecoding simplicityVSAvoidcompression performance
Core Design Contradiction:
Ease of operationVSProductivity

Solution Approach 1:

The patent applies local quality by making the transform type inheritance behavior different for different blocks based on their dimensions. Instead of uniform inheritance for all blocks, the system applies inheritance selectively to blocks that meet specific size conditions, optimizing compression performance for each local block while maintaining a simple decoding process through clear, condition-based rules.

Inventive Principle:
Principle #3Local quality

3Productivity

If transform type information is selectively inherited based on block size, then compression performance is improved, but the number of processing conditions increases

Engineering Contradiction:
Improvecompression performanceVSAvoidnumber of processing conditions
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent applies segmentation by dividing the block space into different categories based on width and height dimensions. By segmenting blocks into those that satisfy inheritance conditions and those that do not, the system creates clear processing paths. This segmentation approach manages complexity by organizing conditions into discrete, easily implementable size-based categories rather than requiring complex multi-parameter decision trees.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentUS20250380004A1Inherited transform type in a merge candidate list
Publication Date: 2025.12.11 TENCENT AMERICA LLC
  • US20250380004A1 patent drawing
  • US20250380004A1 patent drawing
  • US20250380004A1 patent drawing

AI summary

Methods and apparatuses for video decoding and video encoding and a method of processing visual media data are included. The apparatus for video decoding includes processing circuitry configured to receive coded information indicating that a current block is predicted according to a merge mode. When transform type information is to be inherited by the current block, the processing circuitry is configured to add the transform type information of a merge candidate and one of motion information and block vector (BV) information of the merge candidate into a merge candidate list of the current block and reconstruct the current block based on the merge candidate list. The merge candidate list includes the transform type information of the merge candidate and the one of the motion information and BV information of the merge candidate.