Selective Bilateral Filtering for Video Block Parallel Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In video coding, the process of bilateral filtering can hinder parallel processing of neighboring blocks due to dependencies on unfiltered samples, leading to reduced throughput.
Innovation Solution
Implementing a method where video coders selectively filter samples of a current block, refraining from filtering those samples used by neighboring blocks for intra prediction, allowing parallel processing while still benefiting from filtering operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If bilateral filtering is applied to all reconstructed samples, then filtering quality is improved, but parallel processing capability deteriorates due to dependencies on unfiltered samples
Solution Approach 1:
The patent segments the reconstructed block into different sample types: boundary samples that require filtering for quality, and interior samples that can be skipped to enable parallel processing. This segmentation allows selective filtering where it matters most while maintaining processing efficiency.
Solution Approach 2:
The patent applies filtering selectively to specific locations (boundary samples) rather than uniformly across all samples. Boundary samples receive filtering treatment to maintain quality at block edges, while interior samples are processed differently to enable parallel operations.
2Manufacturing precision
If filtering is performed on all samples, then artifact reduction is improved, but processing time increases
Solution Approach 1:
The patent extracts and identifies boundary samples from the reconstructed block that benefit most from filtering. By taking out only these critical samples for filtering operations, the patent reduces overall processing time while maintaining artifact reduction effectiveness at block boundaries.
Solution Approach 2:
The patent applies partial filtering action only to boundary samples rather than excessive filtering of all samples. This partial action is sufficient to reduce artifacts where they occur most frequently (at block boundaries) while avoiding unnecessary processing overhead.
3Productivity
If selective filtering is implemented, then processing efficiency is improved, but filtering consistency deteriorates due to mixed filtered and unfiltered samples
Solution Approach 1:
The patent implements dynamic filtering decisions based on sample location and prediction mode. The filtering operation adapts to different boundary conditions and prediction types, maintaining consistency where needed while allowing flexibility for parallel processing in other regions.
Data Source
AI summary
An example method of filtering a reconstructed block of video data includes obtaining, by one or more processors, reconstructed samples of a current block of the video data; and selectively bilaterally filtering, by the one or more processors, the reconstructed samples of the current block to generate a filtered current block. In this example, selectively bilaterally filtering the reconstructed samples of the current block comprises refraining from bilaterally filtering at least one reconstructed sample of the current block such that the filtered current block includes at least one non-bilaterally filtered sample.