Multiple-Parameter Cross-Component Intra Prediction for Chroma Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies face challenges in efficiently compressing video data while maintaining video quality, particularly in cross-component intra prediction where accurate prediction of chroma samples from luma samples is crucial.
Innovation Solution
The implementation of a cross-component intra prediction (CCIP) mode that uses multiple parameters, specifically a cross-component linear model mode (CCLM) and a convolutional cross-component mode (CCCM), to predict chroma samples based on associated luma samples, utilizing filters with weighing factors to enhance prediction accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If cross-component intra prediction is used to compress video data, then compression efficiency is improved, but prediction accuracy of chroma samples from luma samples deteriorates
Solution Approach 1:
The patent applies parameter changes by using multiple parameters (offset values and gradient values) to adapt the prediction model to different local characteristics of luma-chroma relationships. Different prediction units can have different offset and gradient parameters, allowing the system to adjust prediction accuracy dynamically across different regions while maintaining overall compression efficiency.
Solution Approach 2:
The patent implements local quality by dividing the prediction unit into multiple sub-regions, each with its own offset and gradient parameters. This allows different local areas to have customized prediction characteristics based on their specific luma-chroma relationships, improving prediction accuracy in regions with varying characteristics while maintaining compression efficiency through localized parameter optimization.
2Measurement precision
If multiple parameters are used in cross-component intra prediction, then prediction accuracy is improved, but device complexity increases
Solution Approach 1:
The patent applies preliminary action by pre-calculating offset and gradient parameters during the encoding process and storing them for use during decoding. This allows the decoder to use these pre-computed parameters directly without performing complex calculations, thereby improving prediction accuracy while minimizing the computational complexity actually executed during video playback.
Solution Approach 2:
The patent extracts only the essential parameters (offset and gradient values) needed for accurate prediction and transmits them in the bitstream, separating the complex parameter derivation process from the decoding process. This allows the decoder to use simple parameter application rather than complex computation, improving prediction accuracy while reducing decoding complexity.
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 a syntax element for a cross-component intra prediction mode. A first chroma sample is generated by combining a first luma sample co-located with the first chroma sample and its associated neighboring luma samples using a plurality of weighing factors. The method includes extracting at least one of weighing factors from the video bitstream, and deriving the at least another weighing factor. The method further includes combining the first luma sample and the neighboring luma samples based on the at least one weighing factor that is signaled in the video bitstream and the at least another weighing factor that is not signaled in the video bitstream to generate the first chroma sample, which is applied to reconstruct the current coding block.