In-loop reshaping for video compression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video codec technologies face challenges in achieving optimal compression ratios and complexity reduction, particularly in parallelized implementations, and lack effective tools for in-loop reshaping (ILR) that allow for efficient reconstruction of video units across different domains.
Innovation Solution
The proposed method involves performing a conversion between a coded representation of a video and the actual video, using side information for in-loop reshaping (ILR) that provides parameters for reconstructing video units based on representations in multiple domains, including scaling chroma residue. This method is applicable to existing standards like HEVC and VVC, as well as future standards, and includes specific processes such as obtaining prediction blocks, applying reshaping, and using clipping processes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If in-loop reshaping is applied to improve compression ratio, then compression efficiency is improved, but decoding complexity increases
Solution Approach 1:
The patent applies in-loop reshaping (ILR) as a preliminary processing step within the decoding loop, transforming video units from a first domain to a second domain before final reconstruction. By performing the reshaping transformation early in the decoding process using domain transformation matrices, the patent improves compression efficiency while managing complexity through systematic domain conversion rather than complex post-processing operations.
2Manufacturing precision
If in-loop reshaping is applied to improve reconstruction accuracy, then reconstruction quality is improved, but processing complexity increases
Solution Approach 1:
The patent changes the domain parameters of video units by applying domain transformation matrices that convert representations from a first domain to a second domain. This parameter transformation improves reconstruction accuracy by optimizing the mathematical representation of video data, while the use of standardized matrix operations keeps processing complexity manageable through efficient linear algebra computations.
3Productivity
If domain transformation is applied to improve compression efficiency, then compression performance is improved, but computational complexity increases
Solution Approach 1:
The domain transformation is performed as a preliminary action within the in-loop processing architecture, converting video units to optimized domains before reconstruction and filtering operations. This preliminary transformation improves overall compression efficiency by enabling more effective subsequent processing steps, while the systematic nature of the domain conversion maintains computational tractability.
Solution Approach 2:
The patent employs parameter changes through domain transformation matrices that rerepresent video data in more compression-friendly domains. By changing the mathematical parameters and representation of video units, the system achieves improved compression efficiency while the use of fixed transformation matrices keeps computational complexity bounded and predictable.
Data Source
AI summary
A video processing method is provided to include performing a conversion between a coded representation of a video including one or more video regions and the video. The coded representation includes side information applicable for in-loop reshaping (ILR) of some of the one or more video regions. The side information provides parameters for a reconstruction of a video unit of a video region based on a representation of the video unit in a first domain and a second domain and/or scaling chroma residue of a chroma video unit. The side information applicable to the some of the one or more video regions is coded without directly inheriting from other video regions.