Adaptive Loop Filtering for High Dynamic Range Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding standards lack effective adaptive loop filtering techniques for high-dynamic range (HDR) video, which limits their coding efficiency and image quality.
Innovation Solution
The development of an adaptive loop filtering method that classifies input blocks based on luminance information and computes HDR-specific filtering coefficients, allowing for improved filtering adaptations and bit-depth management, specifically for HDR video using PQ and HLG transfer functions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional video coding standards are used for HDR video, then compatibility with existing systems is maintained, but coding efficiency and image quality are insufficient
Solution Approach 1:
The patent implements dynamic adaptation of loop filtering parameters based on HDR content characteristics. The filtering strength, offset values, and clip levels are adjusted according to the actual luminance range and transfer function (PQ or HLG) of the HDR video content, enabling the system to maintain compatibility while optimizing coding efficiency for different HDR scenarios
Solution Approach 2:
The patent changes key filtering parameters including clip level values, offset ranges, and filtering strength coefficients to be appropriate for HDR luminance ranges. By modifying these parameters based on whether the content uses PQ or HLG transfer functions, the system achieves both compatibility and improved coding performance for HDR video
2Device complexity
If standard loop filtering is applied to HDR video, then processing complexity is low, but image quality and distortion reduction are insufficient
Solution Approach 1:
The patent applies local adaptation by computing filtering parameters based on local luminance statistics of each block or picture region. Different regions with different luminance characteristics receive tailored filtering parameters, improving image quality where needed while maintaining lower complexity in uniform regions
Solution Approach 2:
The patent performs preliminary classification of HDR content type (PQ or HLG) and pre-computes appropriate parameter sets before filtering. This preliminary action enables the main filtering process to proceed with predetermined optimized parameters, balancing quality improvement with processing complexity
3Manufacturing precision
If HDR-specific filtering parameters are used, then image quality improves, but computational overhead and processing complexity increase
Solution Approach 1:
The patent implements partial HDR filtering by applying HDR-specific parameters only to blocks or regions that contain HDR content, while using standard parameters for SDR regions. This selective application reduces computational overhead while maintaining image quality improvements where they are most needed
Solution Approach 2:
The patent performs preliminary detection and classification of HDR content characteristics (transfer function type, luminance range) to pre-select appropriate filtering parameters. This avoids computing multiple parameter sets and reduces real-time computational overhead during the actual filtering process
4Measurement precision
If adaptive classification based on luminance information is implemented, then filtering accuracy improves, but processing time and complexity increase
Solution Approach 1:
The patent segments the luminance range into discrete bands or intervals and assigns predefined filtering parameters to each segment. This segmentation approach maintains filtering accuracy by preserving the adaptive nature of parameter selection while reducing computational complexity through lookup tables and simplified classification logic
Solution Approach 2:
The patent uses simplified parameter adjustment rules based on luminance thresholds rather than complex continuous optimization. By changing parameters discretely based on luminance band classification, the system maintains filtering accuracy while significantly reducing processing time compared to continuous adaptive methods
Data Source
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.