Chroma Intra Prediction Filtering Using Cross-Component Correlation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional digital video coding technologies, such as those based on HEVC and VVC, face challenges in achieving optimal compression efficiency due to reliance on single-sided correlation of neighboring pixels for chroma component prediction, which leads to incomplete utilization of reference information and reduced coding efficiency.
Innovation Solution
The proposed solution refines chroma component prediction by considering the correlation of boundary neighboring pixels in an orthogonal direction, using both upper and left neighboring pixels to improve prediction accuracy and coding efficiency through intra chroma component prediction filtering.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If single-sided correlation of neighboring pixels is used for chroma component prediction, then the prediction process is simple, but the coding efficiency is reduced due to incomplete utilization of reference information
Solution Approach 1:
The patent extends the prediction approach from single-sided (one-dimensional) correlation to double-sided (two-dimensional) correlation by considering both upper and left neighboring pixels. This dimensional expansion allows the system to utilize reference information more completely, thereby improving coding efficiency without excessively increasing complexity
Solution Approach 2:
The patent merges the prediction processes for luma and chroma components by using cross-component prediction. The luma component prediction is combined with chroma component prediction through linear model calculation, allowing the system to leverage the stronger correlation in luma to improve chroma prediction accuracy and coding efficiency
2Measurement precision
If both upper and left neighboring pixels are used for prediction, then prediction accuracy is improved, but the computational complexity increases
Solution Approach 1:
The patent introduces a linear model with parameters alpha and beta to mathematically describe the relationship between luma and chroma components. By changing the prediction approach to use a linear model with adjustable parameters, the system can achieve more accurate predictions while maintaining manageable computational complexity through parameter optimization
Solution Approach 2:
The patent uses the luma component as an intermediary to predict the chroma component. Instead of directly analyzing chroma correlations, the system uses the strongly correlated luma component as a mediator, calculating chroma predictions through the linear model based on luma values, which simplifies the overall computation while improving accuracy
Data Source
AI summary
A picture decoding method comprises: partitioning a picture is partitioned, and a luma component intra prediction mode and a chroma component intra prediction mode of the current coding block are determined; when the chroma component intra prediction mode indicates using a luma component of the current coding block to determine a prediction value of a chroma component of the current coding block, a prediction block of the chroma component of the current coding block is determined according to the luma component intra prediction mode; and prediction refinement is performed on the prediction block of the chroma component of the current coding block, so as to obtain a refined prediction block of the chroma component of the current coding block.