Adaptive Loop Filter Parameter Sets for Picture-Level Video Decoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies face challenges in achieving high compression ratios with minimal sacrifice in picture quality, particularly in scenarios with limited network resources and increasing video quality demands.
Innovation Solution
Implementing adaptive loop filter (ALF) parameter signaling through picture-level syntax elements and parameter set identifiers, ensuring that the temporal identifier of the parameter set is less than or equal to the current picture's identifier, to enable efficient ALF application on luma and chroma components of video slices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If adaptive loop filter parameters are signaled for every picture to maintain picture quality, then picture quality is improved, but bitstream complexity and bandwidth consumption increase
Solution Approach 1:
The patent applies preliminary action by signaling ALF parameters at the picture header level in advance, rather than at the slice level. This allows the decoder to prepare filter parameters before processing slices, reducing the need for repeated parameter signaling and lowering bitstream complexity while maintaining filtering quality across all slices in the picture.
Solution Approach 2:
The patent makes ALF parameters universal by applying them at the picture level rather than slice level. A single set of ALF parameters serves multiple slices within the picture, reducing redundancy. The parameter set identifier references a previously signaled parameter set, allowing the same parameters to be reused across different slices without additional signaling overhead.
2Adaptability or versatility
If ALF parameters are signaled at slice level to enable localized filtering, then filtering adaptability is improved, but syntax complexity and processing overhead increase
Solution Approach 1:
The patent transitions from slice-level to picture-level ALF parameter signaling, changing the dimensional scope of parameter application. By moving up one level in the hierarchy (from slice to picture), the patent reduces the number of parameter sets needed while maintaining the ability to apply filtering across different regions through the picture header structure.
Solution Approach 2:
The patent uses parameter set identifiers to reference previously signaled ALF parameter sets. Instead of repeating full parameter sets for each slice, the decoder copies parameters from the referenced parameter set identifier, reducing syntax complexity while enabling efficient parameter reuse across multiple slices.
3Reliability
If temporal identifier constraints are enforced to ensure decoding order, then decoding reliability is improved, but processing flexibility is reduced
Solution Approach 1:
The patent implements feedback through temporal identifier constraints that ensure parameter sets are decoded in the correct temporal order. The constraint that the parameter set temporal identifier must be less than or equal to the current picture temporal identifier provides feedback control, ensuring that only available and correctly decoded parameters are used, thereby maintaining decoding reliability without requiring complex synchronization mechanisms.
Data Source
AI summary
The present disclosure provides a method of decoding a coded video bitstream, the method comprising: obtaining a picture-level syntax element included in a picture header of the current picture by parsing the video bitstream, wherein the picture-level syntax element is used to specify whether an adaptive loop filter (ALF) is enabled for the current picture; in case that the picture-level syntax element specifies the adaptive loop filter is enabled for the current picture, obtaining a parameter set identifier included in the picture header, wherein the parameter set identifier is used to indicate a first parameter set comprising a set of ALF parameters, and wherein a temporal identifier of the first parameter set is less than or equal to a temporal identifier of the current picture; performing the adaptive loop filter on at least one slice of the current picture based on the first parameter set.