Adaptive Loop Filter Temporal Layer Parameter Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video encoding and decoding systems face challenges in efficiently managing adaptive loop filtering (ALF) parameters across different temporal layers, leading to potential errors and inefficiencies in filtering processes.
Innovation Solution
The proposed solution involves storing ALF parameters in multiple arrays, each corresponding to a specific temporal layer, allowing for the selection of applicable ALF parameters based on the temporal layer of the current region being encoded or decoded.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If a single array is used to store ALF parameters from multiple temporal layers, then memory usage is reduced, but errors occur when parameters are copied from higher temporal layers than the current picture
Solution Approach 1:
The patent divides the single array into multiple separate arrays, with each array dedicated to storing ALF parameters for a specific temporal layer. This segmentation prevents parameters from higher temporal layers from being incorrectly copied to lower temporal layers, thereby resolving the reliability issue while maintaining efficient memory utilization through specialized storage for each layer.
2Device complexity
If ALF parameters are stored without temporal layer differentiation, then device complexity is reduced, but dependencies on higher temporal layers cause filtering errors
Solution Approach 1:
The patent segments the ALF parameter storage structure into temporally differentiated arrays, where each array corresponds to a specific temporal layer. This segmentation introduces minimal complexity overhead while fundamentally preventing cross-layer parameter dependencies that cause filtering errors, thereby significantly improving reliability with acceptable complexity increase.
3Quantity of substance
If a list with indicators to pictures is used instead of a separate ALF array, then memory efficiency is improved, but temporal scalability support is compromised
Solution Approach 1:
The patent segments the parameter storage into multiple arrays organized by temporal layer, which maintains memory efficiency comparable to the list approach while explicitly supporting temporal scalability. Each array is dedicated to a specific temporal layer, enabling proper handling of temporal relationships and scalability requirements that the flattened list approach cannot accommodate.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A video coder may reconstruct a current picture of video data. A current region of the current picture is associated with a temporal index indicating a temporal layer to which the current region belongs. Furthermore, for each respective array of a plurality of arrays that correspond to different temporal layers, the video coder may store, in the respective array, sets of adaptive loop filtering (ALF) parameters used in applying ALF filters to samples of regions of pictures of the video data that are decoded prior to the current region and that are in the temporal layer corresponding to the respective array or a lower temporal layer than the temporal layer corresponding to the respective array. The video coder determines, based on a selected set of ALF parameters in the array corresponding to the temporal layer to which the current region belongs, an applicable set of ALF parameters.