Cross-Component Video Coding With Directional LFNST Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing video coding standards, such as VVC/H.266, do not optimally perform low-frequency non-separable transforms when the block texture contains directionality, as they assume no angular directionality in cross-component prediction, leading to suboptimal compression efficiency.
Innovation Solution
An enhanced method is introduced to determine the intra prediction direction based on texture analysis of neighboring samples, allowing for the selection of an appropriate low-frequency non-separable transform index, thereby improving the correlation between luma and chroma channels.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If the existing mapping function uses transform set index of zero for cross-component prediction, then the implementation is simple, but the compression efficiency deteriorates when the block contains directional texture
Solution Approach 1:
The patent introduces dynamic adaptability by allowing the transform set index to vary based on texture analysis results. Instead of using a fixed transform set index of zero, the system dynamically selects from multiple transform set indexes (0-3) based on the detected texture direction, enabling optimal compression for different texture patterns while maintaining implementation feasibility through predefined transform sets.
Solution Approach 2:
The patent changes the transform set index parameter from a fixed value (0) to a variable that depends on texture analysis. By analyzing the texture direction and mapping it to appropriate transform set indexes, the system optimizes compression efficiency for directional textures while preserving implementation simplicity through the use of predefined transform sets and automated selection logic.
2Device complexity
If a fixed transform set index is used for cross-component prediction, then the processing complexity is low, but the compression efficiency worsens for blocks with directional texture
Solution Approach 1:
The patent performs preliminary texture analysis to determine the appropriate transform set index before applying the transform. By pre-analyzing the texture direction and selecting the optimal transform set in advance, the system avoids complex real-time adjustments during transformation while achieving improved compression efficiency for directional textures.
Solution Approach 2:
The system performs self-service by automatically analyzing its own input data characteristics (texture direction) and selecting the appropriate processing parameters (transform set index) without external intervention. This automated adaptation improves compression efficiency while keeping processing complexity manageable through algorithmic self-determination.
3Ease of operation
If the existing assumption of no angular directionality is made, then the cross-component prediction is simpler, but the transform performance deteriorates when directionality is present
Solution Approach 1:
The patent introduces dynamic adaptability to the prediction process by conditionally applying transform directionality based on texture analysis. When directional texture is detected, the system dynamically adjusts the transform set index to match the texture orientation, improving transform performance while maintaining prediction simplicity for non-directional cases through the default transform set index of zero.
Data Source
AI summary
A method comprising receiving an image block unit of a frame, the image block unit being predicted using a cross-component prediction mode and comprising samples in color channels comprising at least one chrominance channel and one luminance channel (900); determining intra prediction direction based on the samples of said luminance channels of the image block unit, samples in a neighborhood of the luminance channel of the block unit and/or samples in the neighborhood of the block unit (902); selecting a low-frequency nonseparable transform index for low-frequency non-separable transform using the determined intra prediction direction (904); and performing the low-frequency non-separable transform using the selected low-frequency non-separable transform index (906).