Video Block Filtering Across Virtual Boundaries
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding standards face challenges in efficiently handling video unit boundaries and virtual boundaries during the filtering process, leading to suboptimal compression efficiency and complex implementations.
Innovation Solution
The proposed methods involve determining the usage of virtual boundaries and samples for filtering processes based on specific rules related to block dimensions and relationships, applying padding processes, and enabling/disabling sample usage across boundaries to simplify and enhance the filtering operations in video encoding and decoding.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If virtual boundaries are applied during filtering process, then compression efficiency is improved, but implementation complexity increases
Solution Approach 1:
The picture is divided into multiple video units (tiles, slices, bricks) with virtual boundaries between them. The filtering process is segmented to operate independently within each video unit, avoiding cross-boundary filtering operations. This segmentation improves compression efficiency by enabling parallel processing and reducing memory access complexity, while the structured division maintains implementation simplicity through modular design.
Solution Approach 2:
Virtual boundaries act as intermediaries that separate different video units. The filtering process uses these virtual boundaries to determine when to stop filtering operations, preventing access to samples across unit boundaries. This intermediary mechanism simplifies implementation by providing clear stopping criteria for filters without requiring complex cross-boundary coordination.
2Manufacturing precision
If samples across video unit boundaries are used in filtering process, then filtering quality is improved, but processing complexity increases
Solution Approach 1:
The patent extracts and removes the complexity of cross-boundary sample access by disabling the usage of samples across video unit boundaries in the filtering process. Instead of implementing complex logic to handle boundary cases, the solution takes out the problematic cross-boundary operations and processes each video unit independently, maintaining filtering quality within bounds while significantly reducing processing complexity.
Solution Approach 2:
The patent applies preliminary padding processes to generate virtual samples at the boundaries of video units before the filtering operation. This preliminary action ensures that filters have sufficient samples to work with within each video unit without needing to access samples from adjacent units, thereby maintaining filtering quality while avoiding the complexity of cross-boundary sample management.
3Ease of operation
If padding process is applied to generate virtual samples, then filtering operation is simplified, but computational overhead increases
Solution Approach 1:
The patent applies padding process partially, only at the boundaries of video units where it is strictly needed to provide virtual samples for filter operations. Rather than padding the entire picture, the solution applies excessive action only where necessary (at boundaries), simplifying the filtering operation in those specific locations while minimizing overall computational overhead by avoiding unnecessary padding elsewhere.
Data Source
AI summary
A method of video processing includes determining, for a conversion between a picture of a video that includes one or more blocks and a bitstream representation of the video, whether a virtual boundary is enabled for a block within the picture for a filtering process based on a rule related to a relationship between a bottom boundary of the block and the picture. The method also includes performing the conversion based on the determining.