Matrix-Based Intra Prediction for High-Efficiency Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies face challenges in efficiently handling high-resolution video data, particularly in terms of bandwidth demand and coding efficiency, especially with the increasing number of connected devices capable of receiving and displaying video content.
Innovation Solution
Implementing matrix-based intra prediction methods, including affine linear weighted intra prediction (ALWIP) modes, which involve boundary downsampling, matrix vector multiplication, and optional upsampling operations to enhance video coding efficiency in standards like HEVC and VVC.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional intra prediction methods are used in existing video coding standards, then the encoding and decoding processes are simpler, but coding efficiency deteriorates when handling high-resolution video data
Solution Approach 1:
The patent segments the intra prediction process into distinct stages: boundary sample extraction, downsampling to create a reduced boundary, matrix-vector multiplication to generate prediction values, and selective upsampling. This segmentation allows each operation to be optimized independently, improving overall coding efficiency while managing computational complexity through structured processing
Solution Approach 2:
The patent changes the parameters of the prediction process by introducing matrix-based transformation operations and adjustable downsampling/upsampling factors. These parameter changes enable adaptive prediction that scales effectively with high-resolution video, improving coding efficiency by transforming the prediction model to handle larger block sizes more effectively
2Productivity
If matrix-based intra prediction (MIP) techniques are implemented, then coding efficiency improves, but computational complexity increases
Solution Approach 1:
The patent applies partial action by selectively performing upsampling only when needed based on block size criteria. For smaller blocks, the process stops after downsampling and matrix multiplication, avoiding unnecessary upsampling operations. This selective approach maintains coding efficiency benefits while reducing computational complexity in scenarios where full MIP processing is not required
Solution Approach 2:
The patent performs preliminary downsampling of boundary samples before the matrix-vector multiplication operation. By reducing the boundary sample set first, the subsequent matrix operations work with smaller data sets, which reduces the computational burden of the MIP technique while preserving its coding efficiency advantages
3Power
If boundary downsampling is performed in MIP mode, then computational complexity is reduced, but prediction precision may deteriorate
Solution Approach 1:
The patent introduces the downsampled boundary as an intermediary representation that bridges the original high-resolution boundary samples and the final prediction block. This intermediary contains the essential boundary information in a compressed form that enables efficient matrix operations while preserving sufficient detail to maintain prediction precision through the subsequent upsampling process
4Power
If upsampling is selectively applied in MIP mode, then computational complexity is reduced, but prediction accuracy may worsen
Solution Approach 1:
The patent applies partial action by selectively performing upsampling only when needed based on block size criteria. For smaller blocks, the process stops after downsampling and matrix multiplication, avoiding unnecessary upsampling operations. This selective approach maintains coding efficiency benefits while reducing computational complexity in scenarios where full MIP processing is not required
Solution Approach 2:
The patent applies different processing quality levels to different block sizes. Larger blocks that benefit from full MIP processing receive complete treatment including upsampling, while smaller blocks use a simplified process. This local quality approach ensures prediction accuracy is maintained where needed while reducing computational complexity where full precision is less critical
Data Source
AI summary
Devices, systems and methods for digital video coding, which includes matrix-based intra prediction methods for video coding, are described. In a representative aspect, a method for video processing includes performing a conversion between a current video block of a video and a bitstream representation of the current video block according to a rule, where the rule specifies a relationship between samples of the current video block and matrices or offset values applied in a matrix weighted intra prediction (MIP) mode during the conversion, and where the MIP mode includes determining a prediction block of the current video block by performing, on previously coded samples of the video, a boundary downsampling operation, followed by a matrix vector multiplication operation, and selectively followed by an upsampling operation.