Loop Filtering Across Constrained Intra Block Boundaries
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In video coding, constrained intra prediction limits the availability of pixels from neighboring blocks for loop filters like deblocking, sample adaptive offset (SAO), and adaptive loop filters (ALF), particularly in scalable video coding processes such as SHEVC, where pixels from non-intra neighboring blocks are not reconstructed, hindering effective filtering across block boundaries.
Innovation Solution
The proposed techniques involve partial loop filtering and adaptive application of filters, such as SAO and ALF, by using padded pixels or skipping filtering when necessary, and employing asymmetric or symmetric partial filters to restrict the use of unavailable pixels, thereby allowing filtering to proceed even when neighboring blocks are not intra-coded.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If constrained intra prediction is used to improve coding efficiency, then coding efficiency is improved, but loop filtering across block boundaries cannot be performed effectively
Solution Approach 1:
The patent applies partial loop filtering by performing filtering operations only on available pixels within the current block, rather than attempting to filter across the entire block boundary. This partial action allows the filtering process to proceed with the pixels that are actually available, maintaining coding efficiency while still providing filtering benefits where possible.
Solution Approach 2:
The patent implements different filtering behaviors for different regions: intra-predicted blocks receive loop filtering operations while inter-predicted blocks do not. This local differentiation allows the system to maintain coding efficiency for inter blocks while applying filtering where it is most beneficial for intra blocks, resolving the contradiction between coding efficiency and filtering effectiveness.
2Productivity
If pixels from non-intra neighboring blocks are not reconstructed to improve coding efficiency, then coding efficiency is improved, but reference samples for loop filters become unavailable
Solution Approach 1:
The patent extracts and uses only the pixels from the current intra-predicted block for loop filtering operations, excluding the need for pixels from neighboring inter-predicted blocks. This extraction approach allows the filtering to proceed with available data without requiring the reconstruction of pixels from non-intra neighboring blocks, thus maintaining coding efficiency while preserving filtering capability.
Solution Approach 2:
The patent uses padded pixels as an intermediary solution when actual neighboring pixels are unavailable. These padded pixels serve as placeholder reference samples that allow the loop filtering algorithm to execute without requiring the unavailable pixels from inter-predicted blocks, thereby maintaining both coding efficiency and filtering functionality.
3Reliability
If conventional loop filtering is applied across block boundaries to improve video quality, then video quality is improved, but filtering cannot proceed when neighboring blocks are not intra-coded
Solution Approach 1:
The patent performs loop filtering on the portion of the block where pixels are available (the current intra-predicted block) rather than attempting to filter the entire boundary. This partial filtering approach ensures the operation can proceed and improve video quality for the available pixels, while avoiding the infeasibility of requiring pixels from non-intra neighboring blocks.
Solution Approach 2:
The patent uses padded pixels as temporary, disposable placeholders to enable the filtering operation to complete. These padded pixels are not actual image data but serve as temporary fillers that allow the filtering algorithm to execute without failing due to missing reference samples, thus maintaining filtering operation feasibility.
Data Source
AI summary
This disclosure describes techniques for coding video data. In particular, this disclosure describes techniques for loop filtering for video coding. The techniques of this disclosure may apply to loop filtering and/or partial loop filtering across block boundaries in scalable video coding processes. Loop filtering may include, for example, one or more of adaptive loop filtering (ALF), sample adaptive offset (SAO) filtering, and deblocking filtering.