Pairwise Merge Candidate Selection Using Averaged Template Cost

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

In the Enhanced Compression Model (ECM) built on top of Versatile Video Coding (VVC), the calculation of pairwise merge candidates does not consider their template costs effectively, leading to suboptimal selection and reduced coding efficiency.

Innovation Solution

Determine a plurality of pairwise merge candidates based on a sorted list of merge candidates, and select at least one pairwise merge candidate partially based on the averaged template cost of the respective candidates.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If pairwise merge candidates are calculated based on smallest template cost in ECM, then selection simplicity is maintained, but coding efficiency is reduced

Engineering Contradiction:
Improveselection simplicityVSAvoidcoding efficiency
Core Design Contradiction:
Ease of operationVSProductivity

Solution Approach 1:

The patent applies preliminary action by sorting the merge candidate list according to template cost before generating pairwise merge candidates. This pre-ordering allows the selection process to efficiently identify optimal candidates without complex real-time calculations, thereby maintaining selection simplicity while improving coding efficiency through more accurate candidate selection.

Inventive Principle:
Principle #10Preliminary action

2Measurement precision

If multiple pairwise merge candidates are determined for each sorted merge candidate, then candidate selection accuracy is improved, but computational complexity increases

Engineering Contradiction:
Improvecandidate selection accuracyVSAvoidcomputational complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent applies local quality by generating multiple pairwise merge candidates specifically for each merge candidate in the sorted list, rather than uniformly processing all candidates. This targeted approach concentrates computational resources on the most promising candidates (those with lowest template costs), improving selection accuracy while controlling overall computational complexity through selective refinement.

Inventive Principle:
Principle #3Local quality

Data Source

PatentUS20260006219A1On pairwise merge candidates
Publication Date: 2026.01.01 NOKIA TECHNOLOGIES OY
  • US20260006219A1 patent drawing
  • US20260006219A1 patent drawing
  • US20260006219A1 patent drawing

AI summary

An apparatus configured to: determine a plurality of pairwise merge candidates for a current coding unit based on a list of merge candidates sorted according to template cost, wherein multiple pairwise merge candidates are determined for respective merge candidates of the sorted list of merge candidates; and select at least one pairwise merge candidate of the plurality of pairwise merge candidates based, at least partially, on an averaged template cost of the respective pairwise merge candidates of the at least one selected pairwise merge candidate.