3D Video Slice Header Prediction for Error Resilience

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current 3D video coding technologies face challenges in efficiently encoding and decoding video content, particularly in supporting flexible decoding orders of texture and depth view components and predicting syntax elements from slice headers, which affects error resilience and compression efficiency.

Innovation Solution

A slice header prediction method that allows for flexible prediction of syntax elements from earlier slice headers within an access unit, enabling prediction control on a view component basis and grouping syntax elements into Group of Slices (GOS) parameter sets for improved error resilience and compression efficiency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If syntax elements are transmitted in each slice header, then error resilience is improved, but compression efficiency deteriorates due to redundant transmission

Engineering Contradiction:
Improveerror resilienceVSAvoidcompression efficiency
Core Design Contradiction:
ReliabilityVSLoss of information

Solution Approach 1:

The patent merges syntax element transmission across multiple slice headers by introducing a GOS parameter set that groups syntax elements common to multiple slices. Instead of repeating the same syntax elements in each slice header, they are transmitted once in the GOS parameter set and referenced by multiple slices, thereby reducing redundancy while maintaining error resilience through the grouped structure.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The patent segments syntax elements into different categories: those that vary per slice (transmitted in individual slice headers) and those that are common across multiple slices (transmitted in the GOS parameter set). This segmentation allows the system to transmit only necessary information in each slice header while grouping common information separately, improving compression efficiency without sacrificing error resilience.

Inventive Principle:
Principle #1Segmentation

2Device complexity

If fixed decoding order is used for texture and depth view components, then decoding complexity is reduced, but flexibility in supporting different decoding orders deteriorates

Engineering Contradiction:
Improvedecoding complexityVSAvoidflexibility in decoding order
Core Design Contradiction:
Device complexityVSAdaptability or versatility

Solution Approach 1:

The patent introduces dynamic control mechanisms that allow the decoding order of texture and depth view components to be flexibly specified through syntax elements in the slice header and GOS parameter set. The system can adaptively select between different decoding orders (texture-first, depth-first, or interleaved) based on content characteristics and transmission requirements, while the decoder complexity remains manageable through clear syntax specifications and predictable decoding patterns.

Inventive Principle:
Principle #15Dynamics

3Loss of information

If prediction is applied to all syntax elements, then compression efficiency is improved, but control over prediction behavior deteriorates

Engineering Contradiction:
Improvecompression efficiencyVSAvoidcontrol over prediction
Core Design Contradiction:
Loss of informationVSEase of operation

Solution Approach 1:

The patent applies prediction selectively to different categories of syntax elements based on their local characteristics. The GOS parameter set enables separate prediction control for syntax elements that benefit from prediction (such as mode flags and reference picture indices) versus those that should be explicitly transmitted (such as slice boundary information). This local quality approach maintains compression efficiency for predictable elements while preserving control and clarity for elements requiring explicit specification.

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP2810443B1Method and apparatus for video coding
Publication Date: 2021.03.31 NOKIA TECHNOLOGIES OY
  • EP2810443B1 patent drawingFigure 1~2
  • EP2810443B1 patent drawingFigure 3
  • EP2810443B1 patent drawingFigure 4a

AI summary

There is provided a slice header prediction method and apparatuses for 3D video encoding and decoding. In some example embodiments the following features can be derived from the header prediction method. Any decoding order of texture and depth view components may be supported. Also flexible prediction of syntax elements from any slice header appearing earlier in decoding order within the same access unit is allowed. The prediction can be turned on or off on view component basis. The syntax elements of the slice header may be categorized in a few sets and the use of the prediction as well as the prediction source for each set can be individually controlled. By using some example embodiments of the method all syntax elements of the slice header may be predicted.