Adaptive Loop Filtering for High-Dynamic-Range Video

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current video coding standards, such as HEVC, lack effective adaptive loop filtering techniques for high-dynamic-range (HDR) video, which limits coding efficiency and image quality, especially in high-resolution content.

Innovation Solution

The development of an adaptive loop filtering method that classifies input blocks based on luminance information and adjusts filter coefficients for HDR content, using techniques like luminance-based classification, gradient weighting, and separate thresholds for HDR content, to enhance coding efficiency and image quality.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If traditional video coding standards (HEVC) are used without HDR-specific adaptive loop filtering, then device complexity is reduced and ease of operation is maintained, but coding efficiency and image quality deteriorate for HDR content

Engineering Contradiction:
Improvecoding efficiencyVSAvoidfiltering process complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent segments the filtering process by classifying blocks into HDR and SDR types based on luminance values. Different filtering modes are applied to different block types: ALF for HDR blocks and traditional loop filters for SDR blocks. This segmentation allows optimized filtering for HDR content while maintaining compatibility with existing standards for standard content.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces dynamic adaptation of filtering parameters based on block characteristics. The filter decision flag and luminance class values enable the system to dynamically select appropriate filtering modes (ALF or traditional) for each block, optimizing performance for varying content types and luminance ranges.

Inventive Principle:
Principle #15Dynamics

2Manufacturing precision

If HDR-specific adaptive loop filtering with luminance-based classification is implemented, then image quality and coding efficiency improve for HDR content, but device complexity and processing overhead increase

Engineering Contradiction:
Improveimage qualityVSAvoidclassification and filtering complexity
Core Design Contradiction:
Manufacturing precisionVSDevice complexity

Solution Approach 1:

The patent applies local quality optimization by using luminance-based classification to identify HDR blocks and applying ALF specifically to those regions. The luminance class value determines whether a block receives ALF processing, ensuring that enhanced filtering is applied locally where it is most beneficial rather than uniformly across all content.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The patent changes filtering parameters based on luminance characteristics. By computing luminance class values and using them to select filtering modes, the system adapts filter parameters (such as filter coefficients and application decisions) to match the local luminance properties of each block, optimizing image quality for HDR content.

Inventive Principle:
Principle #35Parameter changes

3Productivity

If separate filtering processes are applied for HDR and SDR blocks, then coding efficiency improves and image quality is enhanced, but processing time and computational load increase

Engineering Contradiction:
Improvecompression efficiencyVSAvoidprocessing time
Core Design Contradiction:
ProductivityVSLoss of time

Solution Approach 1:

The patent performs preliminary classification of blocks into HDR and SDR categories using luminance-based criteria before applying filtering. By computing luminance class values in advance and making filter decisions based on these pre-computed values, the system avoids redundant processing and streamlines the filtering workflow, reducing overall processing time.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP3910953A1Adaptive loop filtering for high-dynamic range video
Publication Date: 2021.11.17 DOLBY LABORATORIES LICENSING CORP
  • EP3910953A1 patent drawingFigure 1A
  • EP3910953A1 patent drawingFigure 1B
  • EP3910953A1 patent drawingFigure 2A

AI summary

Methods, processes, and systems are presented for adaptive loop filtering in coding and decoding high dynamic range (HDR) video. Given an input image block, its luminance information may be used to adapt one or more parameters of adaptive loop filtering and compute gradient and directionality information, activity information, a classification index, and adaptive-loop-filtering coefficients.