Adaptive Chroma Prediction Weights for Deviated Luma References
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In video coding standards like H.266/VVC, the prediction accuracy of chroma values is reduced due to deviations in neighboring reference values from the current coding block, leading to inaccuracies in cross-component linear models.
Innovation Solution
A method for predicting video color components that considers correlations and similarities between neighboring reference samples and reconstructed values to determine weight coefficients, constructing a more accurate linear model for chroma prediction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a linear model is constructed using conventional CCLM method considering only neighboring reference values, then the prediction process is simple and fast, but the prediction accuracy of chroma values deteriorates when neighboring reference values deviate greatly from current coding block parameters
Solution Approach 1:
The patent introduces correlation coefficients as new parameters to measure the similarity between neighboring reference blocks and current coding block. These correlation coefficients dynamically adjust the weight of reference values in the linear model, transforming the fixed-parameter conventional approach into a variable-parameter adaptive approach that maintains high prediction accuracy across different scenarios
Solution Approach 2:
The linear model transitions from a static construction method to a dynamic one where the weight of reference values changes based on real-time correlation assessment. The patent dynamically selects and weights reference values according to their correlation with the current block, enabling the model to adapt to varying content characteristics and maintain optimality across different video regions
2Measurement precision
If all neighboring reference samples are equally weighted in the linear model, then the calculation is straightforward, but the prediction accuracy deteriorates when some reference samples are not representative of the current coding block
Solution Approach 1:
The patent applies different weights to different reference samples based on their local correlation quality with the current coding block. High-correlation reference samples receive higher weights while low-correlation samples receive lower weights or are excluded, creating a locally optimized weighting strategy that preserves relevant information and filters out deviated reference values
Solution Approach 2:
The patent introduces a feedback mechanism where correlation coefficients are calculated between reference blocks and current blocks, and this feedback information is used to adjust the weighting of reference values in the linear model. This closed-loop approach ensures that only representative reference samples contribute significantly to the prediction, improving accuracy while reducing the impact of non-representative samples
Data Source
AI summary
A method for reading a bitstream, a method for storing a bitstream and a method for transmitting a bitstream are provided. The method includes: a bitstream is read, and the following operations are executed to decode the bitstream: acquiring a luma component neighboring reference value and a luma component reconstructed value corresponding to a current block, wherein the luma component neighboring reference value is indicative of a luma component parameter corresponding to a neighboring reference sample of the current block; generating a chroma component predicted value corresponding to the current block at least based on a weight coefficient corresponding to the neighboring reference sample related to a position of the current block and the neighboring reference sample; and decoding a video based on the chroma component predicted value.