Separate Luma-Chroma Splitting Trees for Efficient Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing Versatile Video Coding (VVC) standard faces challenges in efficiently partitioning color blocks, particularly for chroma components, leading to increased bit rates and reduced coding efficiency due to the use of shared splitting trees, which do not adequately adapt to the different local characteristics of luma and chroma components.
Innovation Solution
Implementing separate splitting trees for luma and chroma components, allowing each to have its own partitioning structure, and using context formation based on the luma tree to enhance entropy coding efficiency by adapting the context formation for chroma blocks based on luma block characteristics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If shared splitting trees are used for luma and chroma components, then device complexity is reduced, but coding efficiency deteriorates due to inability to adapt to different local characteristics
Solution Approach 1:
The patent divides the previously unified splitting tree structure into separate luma splitting tree and chroma splitting tree. This segmentation allows each color component to have its own independent partitioning structure, enabling adaptation to different local characteristics of luma and chroma components while maintaining manageable complexity through separate processing.
Solution Approach 2:
The patent applies different splitting tree structures to different color components (luma and chroma) based on their specific characteristics. The luma splitting tree and chroma splitting tree can have different partitioning patterns, depths, and configurations, allowing each component to be optimized locally according to its own properties rather than using a uniform approach.
2Productivity
If separate splitting trees are implemented for luma and chroma components, then coding efficiency is improved through better adaptability, but device complexity increases
Solution Approach 1:
The patent performs splitting tree construction and partitioning decisions in advance during the encoding process. The luma and chroma splitting trees are built beforehand based on the respective component characteristics, allowing the decoder to simply follow the pre-determined structure without performing complex real-time decisions, thus managing complexity through preliminary processing.
Solution Approach 2:
The patent introduces separate splitting tree structures as intermediary layers between the raw video data and the compression process. These intermediate structures serve as adaptive partitioning frameworks that bridge the gap between the different characteristics of luma and chroma components and the subsequent compression operations, enabling efficient processing without direct complex interactions between components.
3Adaptability or versatility
If separate splitting trees are used for chroma components, then adaptability to local characteristics is enhanced, but bit rate increases due to additional entropy coding requirements
Solution Approach 1:
The patent employs context formation based on the luma splitting tree to guide the entropy coding of the chroma splitting tree. The luma tree structure provides feedback information about the local characteristics of the video content, which is used to adjust the probability models and coding parameters for chroma. This feedback mechanism allows the system to exploit correlations between luma and chroma to reduce the additional bit rate required for separate chroma partitioning.
Solution Approach 2:
The patent merges the entropy coding processes of luma and chroma by using the luma splitting tree information as context for chroma coding. Instead of completely independent coding of both trees, the system combines the processing by allowing the luma structure to inform and optimize the chroma coding, thereby reducing the total bit rate overhead while maintaining separate partitioning adaptability.
Data Source
AI summary
A method and apparatus for video coding. According to the method, input data associated with a picture area having a first-colour picture area and a second-colour picture area are received, where the input data comprise pixel data for the picture area to be encoded at an encoder side or coded data associated with the picture area to be decoded at a decoder side, and where the first-colour picture area is partitioned into one or more first-colour blocks according to a first-colour splitting tree and the second-colour picture area is partitioned into one or more second-colour blocks according to a second-colour splitting tree. Entropy encoding or decoding is applied to the second-colour splitting tree using context formation, where the context formation has information related to the first-colour splitting tree. The one or more first-colour blocks and the one or more second-colour blocks are then encoded or decoded.