Video Coding Transform Partitioning via CTU Scaling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding technologies face challenges in efficiently determining the maximum transform unit size, which affects transform partitioning and interaction with other coding techniques, leading to hardware complexity and implementation issues in encoder and decoder systems.
Innovation Solution
The proposed method involves determining the maximum transform unit size based on the coding tree unit size, using scaling factors, allowable numbers of transform units, and different maximum sizes for intra and inter coded blocks, to control transform partitioning and improve coding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If the maximum transform unit size is determined independently from CTU size, then flexibility in transform partitioning is improved, but device complexity increases due to additional control parameters and interactions with other coding techniques
Solution Approach 1:
The patent changes the parameter relationship by coupling maximum transform unit size to CTU size through a scaling factor. Instead of treating them as independent parameters, the maximum transform unit size is derived as a function of CTU size (e.g., CTU size divided by a scaling factor like 2 or 4), reducing the number of independent control parameters while maintaining adaptability through the scaling factor.
Solution Approach 2:
The patent makes the CTU size parameter serve multiple functions: it defines the coding tree unit structure and simultaneously determines the maximum transform unit size through the scaling relationship. This multi-functionality reduces the need for separate control mechanisms for transform partitioning, simplifying the overall system complexity.
2Productivity
If different maximum transform sizes are used for intra and inter coded blocks, then coding efficiency is improved, but device complexity increases due to additional conditional logic in encoder and decoder
Solution Approach 1:
The patent applies different maximum transform sizes for different block types (intra vs. inter coded blocks) based on their local characteristics. Intra coded blocks use one maximum transform size while inter coded blocks use another, allowing optimized transform partitioning for each block type's specific requirements without requiring complex conditional logic throughout the entire system.
Solution Approach 2:
The patent uses parameter changes by setting different scaling factors or maximum transform unit sizes based on the block type (intra or inter). This allows the system to adapt the transform partitioning parameters according to the coding mode, improving coding efficiency while keeping the implementation relatively simple through straightforward parameter selection based on block type.
3Device complexity
If transform partitioning is tightly coupled with CTU size, then hardware implementation is simplified, but adaptability in transform unit sizing is reduced
Solution Approach 1:
The patent introduces dynamics by using a scaling factor that can be adjusted to change the relationship between CTU size and maximum transform unit size. This allows the system to adapt the transform unit sizing flexibility while maintaining a relatively simple hardware implementation based on the CTU size, as the scaling factor provides a straightforward mechanism for adjustment without complex logic.
Data Source
AI summary
Aspects of the disclosure provide methods and apparatuses for video encoding/decoding. In some examples, an apparatus for video decoding includes receiving circuitry and processing circuitry. For example, the processing circuitry receives a bitstream including bits corresponding to residues associated with a prediction of a block in a coding tree unit (CTU) and determines a maximum transform size based on the CTU size. The maximum transform size is used in a partition to form, from the residues, one or more transform blocks that are respectively encoded into the bits. Then, the processing circuitry reconstructs based on the maximum transform size and from the bits the residues associated with the prediction of the block, and then the processing circuitry reconstructs samples of the block based on the reconstructed residues and the prediction of the block.