Matrix Intra Prediction Context Coding for Efficient Video Blocks
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 and the need for improved compression methods.
Innovation Solution
Implementing matrix-based intra prediction (MIP) techniques, including affine linear weighted intra prediction (ALWIP), which involves boundary downsampling, matrix vector multiplication, and upsampling operations, to enhance video coding efficiency in standards like HEVC and VVC.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional intra prediction methods are used, then coding simplicity is maintained, but coding efficiency deteriorates for high-resolution video
Solution Approach 1:
The patent segments the prediction process into distinct stages: boundary downsampling to extract key reference samples, matrix vector multiplication for efficient prediction calculation, and selective upsampling for output generation. This segmentation allows each stage to be optimized independently, improving overall coding efficiency while managing computational complexity.
Solution Approach 2:
The patent changes the parameters of the prediction process by using matrix-based operations with configurable dimensions and sampling rates. The boundary downsampling factor and upsampling selection can be adjusted based on block size and content characteristics, allowing adaptive optimization between coding efficiency and computational complexity.
2Device complexity
If boundary downsampling is performed, then computational complexity is reduced, but prediction precision may deteriorate
Solution Approach 1:
The patent introduces an intermediary matrix operation between boundary downsampling and upsampling. This matrix vector multiplication acts as a mediator that processes the downsampled boundary samples to generate prediction values, preserving prediction precision while working with the reduced set of boundary samples, thus maintaining computational efficiency.
Solution Approach 2:
The boundary downsampling is performed as a preliminary action before the main prediction calculation. By pre-processing the boundary samples to extract essential information, the subsequent matrix operations work with fewer data points, reducing computational complexity while the preliminary downsampling preserves the critical spatial relationships needed for accurate prediction.
3Measurement precision
If upsampling operation is always performed, then prediction accuracy is improved, but runtime performance deteriorates
Solution Approach 1:
The patent applies partial upsampling only when necessary based on the block size and prediction mode, rather than always performing upsampling. This selective approach applies the upsampling operation only to the extent needed to achieve adequate prediction accuracy, avoiding unnecessary computational overhead and improving runtime performance while maintaining sufficient prediction quality.
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 encoding a current video block of a video using a matrix intra prediction (MIP) mode in which a prediction block of the current video block is determined 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; and adding, to a coded representation of the current video block, a syntax element indicative of applicability of the MIP mode to the current video block using arithmetic coding in which a context for the syntax element is derived based on a rule.