Adaptive Loop Filtering Using Boundary Strength in Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies face challenges in efficiently managing bandwidth demand and artifact reduction in high-resolution videos, particularly in the context of evolving video coding standards like VVC, where existing in-loop filters struggle to adapt to diverse video content effectively.
Innovation Solution
The use of boundary strength as side information for deblocking filters (DBF-BS) and adaptive loop filters (ALF) is introduced to enhance video coding efficiency by improving filter adaptation and artifact reduction, leveraging boundary strength calculations and adaptive filtering techniques to optimize video processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If existing in-loop filters are used for video coding, then device complexity is reduced, but adaptability to diverse video content deteriorates
Solution Approach 1:
The patent performs preliminary classification of blocks into different types (e.g., intra-prediction blocks, inter-prediction blocks, blocks with specific motion characteristics) before applying filtering. This preliminary action enables the filter to adapt to diverse video content by selecting appropriate filter parameters based on block classification, resolving the contradiction between adaptability and complexity.
Solution Approach 2:
The patent implements dynamic filter parameter adjustment based on block characteristics. Instead of using fixed filter parameters, the filter adapts its parameters (such as filter strength, kernel size) according to the classified block type and local video content properties. This dynamic adaptation improves versatility while maintaining manageable complexity through rule-based parameter selection.
2Manufacturing precision
If stronger filtering is applied to reduce artifacts, then video quality is improved, but bandwidth requirements increase
Solution Approach 1:
The patent applies different filter strengths and types to different regions of the video based on block classification. High-filter-strength regions are applied only where artifacts are present (e.g., block boundaries in inter-prediction blocks), while low or zero filter strength is used in regions that do not require filtering. This local differentiation improves video quality where needed while minimizing the overall bandwidth increase.
Solution Approach 2:
The patent dynamically changes filter parameters (strength, kernel size, type) based on block characteristics and content analysis. By adjusting parameters according to local video properties rather than applying uniform strong filtering, the patent achieves improved video quality with reduced bandwidth requirements compared to blanket strong filtering approaches.
3Object-generated harmful factors
If adaptive filtering parameters are used, then artifact reduction is improved, but computational complexity increases
Solution Approach 1:
The patent performs preliminary block classification using simple criteria (prediction mode, block size, motion characteristics) before applying adaptive filtering. This preliminary classification reduces the computational burden by pre-identifying which blocks require filtering and what type of filtering is appropriate, thereby improving artifact reduction while controlling computational complexity.
Solution Approach 2:
The filter uses information already available in the coded video data (block type, prediction mode, motion vectors) to automatically determine filtering parameters without requiring additional complex analysis. This self-service approach leverages existing data structures to guide adaptive filtering, improving artifact reduction while avoiding excessive computational complexity.
Data Source
AI summary
A mechanism for processing video data is disclosed. The mechanism includes determining to employ a boundary strength of a deblocking filter (DBF-BS) as side information input into an adaptive loop filter (ALF) or a cross component ALF (CC-ALF). A conversion can then be performed between a visual media data and a bitstream based on the ALF or CC-ALF.