Cross-Component Linear Models for Luma-Based Chroma Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video signal processing methods lack efficiency in encoding and decoding, particularly in predicting chroma components based on luma components.
Innovation Solution
A method and apparatus for video signal processing that involves down-sampling reconstructed luma components, acquiring maximum and minimum luma values, and using these to predict chroma values through a cross-component linear model.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional video compression techniques are used, then basic compression is achieved, but coding efficiency is insufficient
Solution Approach 1:
The patent introduces a cross-component linear model as an intermediary mechanism that establishes a mathematical relationship between luma and chroma components. This model acts as a mediator to predict chroma values based on luma values, thereby improving chroma prediction accuracy without directly processing chroma data, which resolves the contradiction between coding efficiency and chroma prediction accuracy.
Solution Approach 2:
The patent changes the parameter representation by using linear model parameters (alpha and beta coefficients) to describe the relationship between luma and chroma components. By transforming the prediction problem into a parameter estimation problem, the system achieves better coding efficiency while maintaining high prediction accuracy through optimized parameter selection and transmission.
2Measurement precision
If chroma components are independently encoded, then encoding simplicity is maintained, but prediction accuracy deteriorates
Solution Approach 1:
The patent segments the video encoding process into distinct luma processing and chroma prediction stages. By separating the complex chroma prediction task from the main encoding flow and handling it through a dedicated linear model, the system improves chroma prediction accuracy while keeping the overall encoding structure manageable and modular.
3Productivity
If more compression techniques are applied, then compression ratio increases, but processing complexity increases
Solution Approach 1:
The patent performs preliminary action by establishing the cross-component linear model during the encoding phase and transmitting the model parameters in advance. This allows the decoder to use pre-computed parameters for chroma prediction, reducing the processing complexity at decoding while maintaining high compression efficiency through accurate prediction.
Data Source
AI summary
The video signal processing method comprises the steps of: downsampling luma components of reconstructed blocks adjacent to a current block; acquiring a maximum luma value from among the downsampled luma components based on a first index; acquiring a next-highest maximum luma value from among the downsampled luma components based on a second index; acquiring a maximum average luma value based on an average value of the maximum luma value and the next-highest maximum luma value; acquiring a next-lowest minimum luma value from among the downsampled luma components based on a third index; acquiring a minimum luma value from among the downsampled luma components based on a fourth index; and acquiring a minimum average luma value based on the next-lowest minimum luma value and the minimum luma value.