Unified Clipping for Illumination Compensation in Multi-View Video

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current multi-view video coding techniques face inefficiencies due to different clipping operations for uni-prediction and bi-prediction when illumination compensation is enabled, leading to potential mismatches and increased computational complexity.

Innovation Solution

A unified clipping process is introduced for both uni-prediction and bi-prediction, ensuring identical prediction outputs regardless of the prediction method, and allowing for the conversion of bi-prediction to uni-prediction when motion information refers to the same reference block, thereby simplifying the derivation of aligned inter-view prediction.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If different clipping operations are applied for uni-prediction and bi-prediction when illumination compensation is enabled, then the prediction outputs may differ according to the prediction method, but this leads to mismatches and increased computational complexity

Engineering Contradiction:
Improveprediction output consistencyVSAvoidclipping operation complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent merges the previously separate clipping operations for uni-prediction and bi-prediction into a single unified clipping process. When illumination compensation is enabled, both prediction methods now apply the same clipping operation to their respective reference blocks, eliminating mismatches and reducing computational complexity while maintaining prediction output consistency.

Inventive Principle:
Principle #5Merging (Combining)

2Measurement precision

If bi-prediction is used with illumination compensation, then prediction accuracy may be improved, but the computational complexity and processing time increase

Engineering Contradiction:
Improveprediction accuracyVSAvoiddecoding time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent applies the unified clipping process to both uni-prediction and bi-prediction paths, allowing for efficient implementation. When motion information refers to the same reference block, the system can convert bi-prediction to uni-prediction, reducing computational complexity and processing time while maintaining prediction accuracy through the consistent clipping operation.

Inventive Principle:
Principle #5Merging (Combining)

3Productivity

If a unified clipping process is applied to both uni-prediction and bi-prediction, then computational complexity is reduced and mismatches are eliminated, but the flexibility to handle different prediction methods differently is lost

Engineering Contradiction:
Improvecoding efficiencyVSAvoidprediction method flexibility
Core Design Contradiction:
ProductivityVSAdaptability or versatility

Solution Approach 1:

The patent creates a universal clipping process that serves both uni-prediction and bi-prediction methods. This unified approach maintains flexibility by allowing the system to choose between prediction methods while applying consistent clipping rules, thereby improving coding efficiency without sacrificing the ability to adapt to different prediction scenarios.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Data Source

PatentEP2923491B1Method and apparatus for bi-prediction of illumination compensation
Publication Date: 2018.10.31 HFI INNOVATION INC
  • EP2923491B1 patent drawingFigure 1
  • EP2923491B1 patent drawingFigure 2A~2B
  • EP2923491B1 patent drawingFigure 3

AI summary

A method and apparatus for deriving aligned inter-view prediction for uni-prediction and bi-prediction in three-dimensional and multi-view coding with illumination compensation enabled are disclosed. When the illumination compensation is enabled, the derivation process for a single reference block in reference list O or reference list 1 in the case of uni-prediction is aligned with the derivation process for a reference block in reference list O and a reference block in reference list 1 in the case of bi-prediction. The reconstruction process generates the reference block based on a temporal reference block when Inter prediction is used and based on an inter-view reference block when inter-view prediction is used. For the uni-prediction and bi-prediction, the same clipping process can be included in the reconstruction process, the illumination compensation process, or both.