Chroma Block Prediction Using Reconstructed Luma Samples
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding techniques face challenges in efficiently utilizing neighboring reconstructed samples of the current chroma block and corresponding luma block for improved video coding efficiency and quality, particularly in high-resolution and high-frame-rate scenarios.
Innovation Solution
A method and apparatus that reconstruct a current chroma block by using prediction information of a corresponding luma block, leveraging the reconstructed sample values and neighboring samples to model relationships between the luma and chroma components, enabling efficient intra prediction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If existing intra prediction modes (planar, DC, horizontal, vertical, DM, CCLM) are used for chroma component prediction, then video coding efficiency is improved, but the utilization of neighboring reconstructed samples of current chroma block and corresponding luma block is insufficient
Solution Approach 1:
The patent segments the prediction process into multiple stages: first predicting the luma block using conventional modes, then using the reconstructed luma block samples to predict the chroma block through cross-component linear model. This segmentation allows efficient utilization of neighboring reconstructed samples from both luma and chroma blocks independently, resolving the contradiction between coding efficiency and sample utilization adaptability.
Solution Approach 2:
The patent introduces the reconstructed luma block as an intermediary element that bridges the prediction of chroma block. By using the reconstructed luma block samples as reference for chroma prediction, the system enables more effective utilization of neighboring reconstructed samples while maintaining high coding efficiency through the cross-component linear model.
2Manufacturing precision
If image size, resolution, and frame rate are increased to improve video quality, then video quality is enhanced, but the amount of data to be encoded increases
Solution Approach 1:
The patent merges the prediction processes of luma and chroma components by using the reconstructed luma block to predict the chroma block. This merging approach allows simultaneous compression of both components with shared prediction resources, reducing the total data量 while maintaining high video quality at increased resolution and frame rates.
Solution Approach 2:
The patent changes the prediction parameters by introducing cross-component linear model that utilizes reconstructed luma block samples as reference for chroma prediction. This parameter change enables more efficient encoding of chroma components, reducing the data quantity while preserving video quality improvements from higher resolution and frame rates.
Data Source
AI summary
A method and an apparatus are disclosed for video coding using luma component-based chroma component prediction. In the disclosed embodiments, a video decoding device derives a corresponding luma block to the current chroma block based on a color format. The video decoding device derives a reconstructed region of a luma component for the corresponding luma block and derives a reconstructed region of a chroma component for the current chroma block based on a block partitioning structure of the luma component and the chroma component and prediction information of the corresponding luma block. The video decoding device models a relationship between samples in the reconstructed region of the luma component and samples in the reconstructed region of the chroma component, and generates a prediction block of the current chroma block from samples in the corresponding luma block by using the modeled relationship.