Cross-Component Intra Prediction for Video Color Compression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies struggle to effectively utilize spatial correlations between different color components in video data, leading to inefficiencies in compression and quality retention, particularly in lossy compression scenarios.
Innovation Solution
Implement cross-component intra prediction (CCIP) mode where samples of a first color component are determined based on a linear or non-linear function using co-located and adjacent samples of a second color component, with parameters derived from neighboring areas or explicitly signaled in the video bitstream, compensating for spatial misalignments caused by camera artifacts or post-processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional video coding standards (HEVC, VVC, AV1) are used, then video data can be compressed for transmission and storage, but spatial correlations between different color components are not effectively utilized, leading to suboptimal compression efficiency and quality retention
Solution Approach 1:
The patent merges the prediction of different color components by establishing cross-component prediction relationships. Specifically, it predicts chroma component samples using luma component samples through linear or non-linear functions, combining information from multiple color components to improve prediction accuracy and reduce redundancy between components.
Solution Approach 2:
The patent introduces an intermediate prediction mechanism where luma samples serve as mediators to predict chroma samples. The system uses co-located and adjacent luma samples, applying linear or non-linear transformation functions as intermediaries to generate chroma predictions, thereby effectively transferring spatial correlation information between color components.
2Quantity of substance
If lossy compression is applied to reduce data volume, then bandwidth and storage requirements are reduced, but video quality degradation occurs
Solution Approach 1:
The patent applies preliminary prediction actions before actual encoding. By predicting chroma samples from luma samples using cross-component intra prediction, the system pre-processes the data to exploit inter-component correlations, reducing the residual information that needs to be encoded and transmitted, thereby improving compression efficiency before the lossy compression stage.
3Productivity
If cross-component intra prediction is implemented to improve prediction accuracy, then compression efficiency is enhanced, but computational complexity increases
Solution Approach 1:
The patent applies local quality by using co-located and adjacent luma samples specifically for predicting chroma samples in the same or neighboring blocks. The prediction is performed locally within coding units, using only the necessary neighboring samples and gradient calculations, rather than processing the entire image, thereby limiting computational complexity to local regions while maintaining high prediction accuracy.
Data Source
AI summary
The various implementations described herein include methods and systems for coding video. In one aspect, a video bitstream includes a current coding block of an image frame and includes a cross-component intra prediction mode. A computing system identifies a sample of the first color component and a sample of the second color component co-located with the sample of the first color component. At least two adjacent samples of the second color component are identified, and a location of each adjacent sample is identified by a horizontal delta coordinate value or a vertical delta coordinate value with respect to the sample of the second color component. The computing system generates the sample of the first color component based on the at least two adjacent samples of the second color component, and reconstructs the current coding block based at least on the generated sample of the first color component.