Video Bitstream Prediction Using Weighted Template Partitioning

Solution Overview

Problem

Existing video coding technologies suffer from inaccurate prediction modes due to insufficient template partitioning, leading to poor compression efficiency.

Innovation Solution

A method is introduced to determine a weight derivation mode for a current block, which allows for determining at least one of K prediction modes based on the weight of a template, where K is a positive integer greater than 1, to improve prediction accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If template partitioning is used to determine prediction modes, then prediction can be performed, but the partitioning is not fine enough leading to inaccurate prediction modes

Engineering Contradiction:

Improveprediction mode accuracyVSAvoidtemplate partitioning complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The template is divided into multiple sub-templates through fine-grained partitioning. The decoder divides the template into a first sub-template and a second sub-template based on a partition offset, allowing separate prediction mode determination for each sub-template. This segmentation enables more accurate local prediction while maintaining manageable complexity through systematic division.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

Different prediction modes are determined for different sub-templates based on their local characteristics. The partition offset is calculated specifically for each sub-template to capture local temporal and spatial redundancies. This local quality approach ensures that each region is predicted with the most appropriate mode for its specific characteristics, improving overall prediction accuracy.

Inventive Principle:

Principle #3Local quality

2Measurement precision

If more prediction modes are determined using template matching, then prediction accuracy improves, but the computational complexity increases

Engineering Contradiction:

Improveprediction mode accuracyVSAvoidencoding/decoding efficiency

Core Design Contradiction:

Measurement precisionVSProductivity

Solution Approach 1:

The template is divided into multiple sub-templates through fine-grained partitioning. The decoder divides the template into a first sub-template and a second sub-template based on a partition offset, allowing separate prediction mode determination for each sub-template. This segmentation enables more accurate local prediction while maintaining manageable complexity through systematic division.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

Instead of performing exhaustive template matching across the entire template, the method performs partial matching on divided sub-templates. This partial action approach determines prediction modes for only the necessary sub-regions, reducing computational complexity while maintaining sufficient prediction accuracy for the current block.

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS12526401B2Method for video encoding/decoding and bitstream transmission

Publication Date: 2026.01.13 GUANGDONG OPPO MOBILE TELECOMMUNICATIONS CORP LTD

AI summary

Methods for video encoding/decoding and bitstream transmission are provided. A weight of a template is determined according to a weight derivation mode. At least one number of K prediction modes for a current block is determined. A prediction value is determined according to the K prediction modes and the weight derivation mode. That is, in the disclosure, a template is simulately partitioned according to the weight.