Matrix-Weighted Intra Prediction for Video Coding Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding systems, such as VVC, face challenges in efficiently handling various video sources including 3D signals and maintaining optimal video quality due to impairments during encoding and decoding processes, particularly in intra prediction modes.
Innovation Solution
The method introduces Matrix Weighted Intra Prediction (MIP) mode, which utilizes a candidate mode group including a Matrix-based Intra Prediction (MIP) mode to determine a target prediction mode for current blocks, incorporating transform sets, transpose flags, and transform kernels, and employs histogram/gradient/distortion analysis to select pre-defined intra prediction modes based on block size and syntax elements, enhancing intra prediction performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional intra prediction modes are used, then the coding process is simple, but the video quality deteriorates due to impairments during encoding and decoding
Solution Approach 1:
The patent applies parameter changes by transforming the intra prediction process through matrix operations. Instead of using traditional angular prediction modes directly, the patent transforms reference pixel values using predefined matrices to generate prediction values, thereby improving video quality while managing complexity through standardized matrix transformations.
Solution Approach 2:
The patent introduces matrix transformations as an intermediary step between reference pixels and prediction values. This intermediary transformation process enhances prediction accuracy by applying mathematical operations that compensate for encoding/decoding impairments, thereby improving video quality without directly modifying the core prediction architecture.
2Productivity
If Matrix-based Intra Prediction (MIP) mode is introduced to improve video quality, then the coding efficiency improves, but the computational complexity increases
Solution Approach 1:
The patent segments the intra prediction process into distinct stages: reference pixel selection, matrix transformation, and prediction value generation. This segmentation allows for optimized implementation of each stage, improving coding efficiency while managing computational complexity through modular processing.
Solution Approach 2:
The patent employs universal matrix transformations that can be applied across different block sizes and prediction directions. By using a unified matrix-based approach rather than separate prediction algorithms for each case, the patent improves coding efficiency through consistent processing while reducing overall computational complexity.
3Reliability
If adaptive loop filtering is applied to reconstructed video data, then the video quality improves, but the processing time and complexity increase
Solution Approach 1:
The patent applies preliminary action by performing matrix-based predictions before the final reconstruction and filtering stages. This advance prediction with built-in transformation compensates for potential impairments early in the process, reducing the burden on subsequent adaptive loop filtering operations and thereby decreasing overall processing time while maintaining video quality.
Data Source
AI summary
A method and apparatus for matrix weighted intra prediction. According to this method, input data associated with a current block are received, wherein the input data comprise pixel data to be encoded at an encoder side or coded data associated with the current block to be decoded at a decoder side, and wherein a target prediction mode is determined from a candidate mode group for the current block and the candidate mode group comprises a Matrix-based Intra Prediction (MIP) mode. A traditional intra prediction mode is determined from a traditional intra mode group, wherein the traditional intra mode group comprises multiple angular intra prediction modes, wherein the traditional intra prediction mode is used to process the current block, a subsequent block, and/or a collocated block of the current block. The current block is encoded or decoded using the target prediction mode.