3D Video Codec Flags for Reliable Inter-Component Prediction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing 3D video codecs face issues with inter-component prediction due to the presence or absence of reference component layers in the bitstream, leading to incorrect enablement or disablement of prediction tools, which affects the decoding process.

Innovation Solution

Implement conditional signaling of flags in the data stream to control the enablement/disablement of inter-component prediction for coding units, ensuring that prediction is only applied if all required reference layers are present, and allowing for flexible enablement/disabling based on the availability of depth and texture layers.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If inter-component prediction is enabled based on the presence of reference component layers in the bitstream, then prediction tools can be applied more broadly, but incorrect prediction may occur when required pictures are missing

Engineering Contradiction:
Improveapplicability of inter-component predictionVSAvoidcorrect operation of prediction
Core Design Contradiction:
Adaptability or versatilityVSReliability

Solution Approach 1:

The patent applies local quality by making the enablement of inter-component prediction specific to each coding unit (CU) rather than uniformly across the entire picture. The flag in_comp_pred_flag is evaluated separately for each CU based on the availability of reference pictures required by that specific CU, allowing different regions of the picture to have different prediction enablement states according to their local reference picture availability.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The patent introduces dynamics by making the inter-component prediction enablement flexible and adaptive rather than static. The system dynamically determines whether to apply inter-component prediction for each CU based on real-time assessment of reference picture availability, allowing the prediction mechanism to adapt its behavior to the actual state of reference pictures in the bitstream.

Inventive Principle:
Principle #15Dynamics

2Reliability

If inter-component prediction is disabled when any reference layer is missing, then prediction accuracy is maintained, but prediction is overly restricted even when some reference layers are available

Engineering Contradiction:
Improveprediction accuracyVSAvoiddecoding efficiency
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The patent enables different parts of the picture to have different prediction enablement states based on their specific reference picture requirements. Each CU is evaluated independently, so that CUs with available reference pictures can utilize inter-component prediction while CUs with missing references can fall back to other prediction methods, optimizing overall decoding efficiency.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The patent segments the picture into multiple coding units (CUs) and applies inter-component prediction enablement independently to each CU. This segmentation allows the system to apply prediction only where reference pictures are available, rather than applying a uniform enablement/disabling decision to the entire picture, thereby improving overall decoding productivity.

Inventive Principle:
Principle #1Segmentation

3Loss of information

If a single flag controls inter-component prediction for the entire picture, then signaling overhead is reduced, but flexibility to apply prediction selectively is lost

Engineering Contradiction:
Improvesignaling overheadVSAvoidselective prediction capability
Core Design Contradiction:
Loss of informationVSAdaptability or versatility

Solution Approach 1:

The patent implements a self-service mechanism where each coding unit independently determines whether inter-component prediction should be applied based on its own reference picture availability and the flag value. The system uses the flag in_comp_pred_flag to enable or disable prediction on a per-CU basis, allowing selective application without requiring separate flags for each CU, thus balancing signaling overhead with adaptability.

Inventive Principle:
Principle #25Self-service

Data Source

PatentEP4672754A13d-video codec supporting inter-component prediction
Publication Date: 2025.12.31 FRAUNHOFER GESELLSCHAFT ZUR FORDERUNG DER ANGEWANDTEN FORSCHUNG EV
  • EP4672754A1 patent drawingFigure 1
  • EP4672754A1 patent drawingFigure 2
  • EP4672754A1 patent drawingFigure 3

AI summary

3D video coding/decoding supporting inter-component prediction is made operatable with respect to a broader range of settings with respect to accompanying individual views with depth and deactivation/enablement of inter-component prediction, by using one or more flags in the data stream to control the enablement/disablement of the availability of inter-component prediction for coding units of a current picture.