Weighting Factors for Prediction Sample Filtering in Intra Modes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding technologies face challenges in efficiently handling non-square blocks and wide-angle intra prediction modes, particularly in determining weighting factors for neighboring samples in planar and DC modes, which affects prediction accuracy and bandwidth usage.
Innovation Solution
The proposed method involves deriving weighting factors for neighboring samples based on the dimensions and positions of the current video block, using a position-dependent intra prediction combination (PDPC) method that combines neighboring samples with a prediction signal to generate a refined prediction signal, specifically for planar and DC modes, and applying these factors during video encoding and decoding processes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional intra prediction methods are used for non-square blocks, then the coding process is simple, but the prediction accuracy is insufficient
Solution Approach 1:
The patent applies different weighting factor determination methods to different prediction modes (planar/DC modes use dimension/position-based weighting, while angular modes use gradient-based weighting). This local differentiation optimizes prediction accuracy for each mode type without uniformly increasing complexity across all modes, resolving the contradiction between accuracy and complexity.
2Measurement precision
If position-dependent intra prediction combination is applied to all modes, then prediction accuracy improves, but computational complexity increases
Solution Approach 1:
The patent segments the intra prediction modes into two categories: planar/DC modes that benefit from position-dependent weighting factors based on block dimensions, and angular modes that use gradient-based weighting. This segmentation allows selective application of computationally intensive methods only where beneficial, improving accuracy while controlling overall complexity.
Solution Approach 2:
The patent applies position-dependent intra prediction combination selectively to planar and DC modes rather than all intra prediction modes. This partial application provides sufficient accuracy improvement for modes where it matters most while avoiding unnecessary computational overhead in angular modes, resolving the power-complexity tradeoff.
3Measurement precision
If weighting factors are determined from block dimensions and positions, then prediction quality improves for planar and DC modes, but the method becomes more complex
Solution Approach 1:
The patent changes the parameters used for weighting factor determination based on prediction mode: for planar and DC modes, it uses block dimensions and sample positions as parameters; for angular modes, it uses gradient calculations. This parameter adaptation improves prediction quality for each mode type while keeping the overall method manageable through systematic parameter selection.
Data Source
AI summary
A video processing method is provided to include: deriving weighting factors for neighboring samples of samples of a current video block of a video according to a rule; and performing a conversion between the current video block and a coded representation of the video, and wherein the rule specifies that, in case that the current video block is coded using a planar mode or a DC mode, the weighting factors are determined from at least one of a dimension of the current video block or positions of the samples of the current video block, wherein the current video block uses a PDPC method that combines the neighboring samples with a prediction signal of the current video block to generate a refined prediction signal, and wherein a weighting factor of a prediction signal of a sample is determined based on weighting factors of corresponding neighboring samples of the sample.