Multi-view Video Illumination Compensation Block-Level Encoding

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Multi-view Video Coding systems with heterogeneous cameras face challenges in illumination and color compensation, leading to decreased efficiency in cross-view prediction due to luminance and chrominance differences across different camera views, which existing methods inadequately address, especially when illumination mismatches are significant.

Innovation Solution

The implementation of illumination and color compensation methods within the Multi-view Video Coding framework, utilizing new syntax elements such as ic_prediction_flag, ic_enable, and ic_sym for illumination compensation, and cc_offset for color compensation, allowing for differential encoding and quantization of illumination and color offsets, and enabling switchable compensation on a block basis to address varying levels of mismatch.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If illumination compensation parameters are applied for each block, then prediction accuracy is improved, but coding overhead increases

Engineering Contradiction:
Improveprediction accuracyVSAvoidcoding overhead
Core Design Contradiction:
Measurement precisionVSQuantity of substance

Solution Approach 1:

The patent applies illumination compensation selectively at the block level rather than globally. Each block can independently enable illumination compensation based on local illumination mismatch characteristics, allowing high prediction accuracy in regions with significant illumination differences while avoiding unnecessary overhead in regions without such differences.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The patent divides the picture into multiple blocks and applies illumination compensation parameters (scale and offset) independently to each block. This segmentation allows the system to adapt to local illumination variations without applying compensation everywhere, thus balancing prediction accuracy with coding efficiency.

Inventive Principle:
Principle #1Segmentation

2Productivity

If color compensation is applied to correct chrominance differences, then cross-view prediction efficiency is improved, but device complexity increases

Engineering Contradiction:
Improvecross-view prediction efficiencyVSAvoiddevice complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent introduces color compensation parameters (cc_offset) to adjust chrominance values across different views. By modifying the color parameters rather than reprocessing the entire video data, the system improves cross-view prediction efficiency while keeping the additional complexity manageable through parameter-based correction.

Inventive Principle:
Principle #35Parameter changes

3Ease of manufacture

If global illumination compensation is used, then coding simplicity is maintained, but local illumination mismatches are not adequately addressed

Engineering Contradiction:
Improvecoding simplicityVSAvoidillumination compensation accuracy
Core Design Contradiction:
Ease of manufactureVSMeasurement precision

Solution Approach 1:

The patent transitions from global illumination compensation to block-level local compensation. Each block can have its own illumination compensation parameters, allowing the system to maintain coding simplicity through a standardized block-based approach while significantly improving accuracy by adapting to local illumination conditions.

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP2424244B1Apparatus for illumination and color compensation for multi-view video decoding
Publication Date: 2019.09.11 DOLBY INTERNATIONAL AB
  • EP2424244B1 patent drawingFigure 1
  • EP2424244B1 patent drawingFigure 2
  • EP2424244B1 patent drawingFigure 3

AI summary

There are provided a method and apparatus for illumination and color compensation for multi-view video coding. A video encoder includes an encoder (100) for encoding a picture by enabling color compensation of at least one color component in a prediction of the picture based upon a correlation factor relating to color data between the picture and another picture. The picture and the other picture have different view points and both corresponding to multi-view content for a same or similar scene.