Cross-Component Geometric Partitioning for Video Motion Prediction

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing video coding technologies use predefined straight-line partitioning boundaries that are sub-optimal for irregular video objects, leading to inefficient motion prediction and video encoding/decoding accuracy.

Innovation Solution

Implement geometric partitioning modes that allow for more flexible partitioning boundaries, using geometric partitions and blending techniques to improve motion prediction accuracy for video encoding and decoding.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If predefined straight-line partitioning boundaries are used, then device complexity is reduced and ease of operation is improved, but motion prediction accuracy and video encoding quality deteriorate

Engineering Contradiction:

Improveease of partitioning operationVSAvoidmotion prediction accuracy

Core Design Contradiction:

Ease of operationVSMeasurement precision

Solution Approach 1:

The video block is divided into multiple partitions using geometric partitioning modes (GPM) that allow different partitioning patterns for different color components. The luma block and chroma blocks are segmented independently, enabling each to use partitioning boundaries optimized for its specific characteristics, thus improving motion prediction accuracy while maintaining operational simplicity through systematic segmentation rules

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent extends the partitioning approach from a single straight-line boundary to multiple geometric partitioning patterns by adding dimensional complexity. Different geometric partitions (e.g., horizontal, vertical, diagonal splits) are applied to different color components, transforming the simple 1D partitioning into a multi-dimensional partitioning strategy that better captures irregular video object shapes

Inventive Principle:

Principle #17Another dimension (Dimensionality change)

2Device complexity

If single straight line partitioning boundary is used, then device complexity is reduced, but the ability to represent irregular video object shapes deteriorates

Engineering Contradiction:

Improvepartitioning structure complexityVSAvoidadaptability to irregular video objects

Core Design Contradiction:

Device complexityVSAdaptability or versatility

Solution Approach 1:

Different partitioning patterns are applied locally to different color components based on their specific needs. The luma component may use one geometric partitioning pattern while chroma components use different patterns, allowing each region and component to have the partitioning quality optimized for its local characteristics and irregular shapes

Inventive Principle:

Principle #3Local quality

Solution Approach 2:

The partitioning structure becomes dynamic by allowing different geometric partitioning modes to be selected and applied to different color components. The system can adaptively choose from multiple partitioning patterns (horizontal, vertical, diagonal) depending on the content characteristics, making the partitioning structure flexible and adaptable to various irregular video object shapes

Inventive Principle:

Principle #15Dynamics

3Measurement precision

If geometric partitioning modes with flexible boundaries are implemented, then motion prediction accuracy is improved, but device complexity and computational requirements increase

Engineering Contradiction:

Improvemotion prediction accuracyVSAvoidencoding/decoding complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent applies geometric partitioning selectively to specific color components (luma and chroma) rather than uniformly to all blocks. By applying the complex partitioning only where needed (inter-blocks with geometric partitioning mode) and using simpler partitioning elsewhere, the system achieves partial action that improves accuracy while controlling overall complexity

Inventive Principle:

Principle #16Partial or excessive action

Solution Approach 2:

The system changes partitioning parameters (partitioning mode, boundary orientation, split position) based on the specific characteristics of each color component and block type. By dynamically adjusting these parameters rather than using fixed partitioning, the system achieves better motion prediction accuracy while managing complexity through parameter optimization rather than structural complexity

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS12445620B2Systems and methods for cross-component geometric/wedgelet partition derivation

Publication Date: 2025.10.14 TENCENT AMERICA LLC

AI summary

The various implementations described herein include methods and systems for coding video. In one aspect, a method includes receiving video data including a picture, wherein the picture is coded using at least a first color component and a second color component, and the picture includes a first block that is coded in a geometric partition mode, the first block including a first geometric partition and a second geometric partition; reconstructing samples in a first geometric partition of the first color component of the first block; deriving samples in the first geometric partition of the second color component of the first block based on the reconstructed samples of the first color component of the first block; and decoding the first block in the picture based at least on the reconstructed samples in the first geometric partition of the first color component and the second color component of the first block.