Chroma Block Reconstruction Using Multi-Model Linear Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The challenge of efficiently encoding and decoding high-definition and ultra-high-definition video data while maintaining image quality is exacerbated by the exponential growth in video data volume, necessitating improved methods for video compression and decompression.
Innovation Solution
Implementing a multi-model linear model (MMLM) to predict chroma components based on decoded luma components within a coding unit, utilizing a threshold luma value to generate a multi-model linear model for reconstructing chroma blocks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional video compression standards are used, then video data can be encoded and decoded, but the exponential growth in video data volume makes efficient compression increasingly difficult
Solution Approach 1:
The patent segments the chroma prediction process into multiple linear models based on different luma value ranges. Instead of using a single prediction model for the entire chroma block, the method divides the prediction task into multiple segments, each handled by a dedicated linear model. This segmentation allows more precise local prediction adapted to different brightness regions, improving compression efficiency without requiring more bits.
Solution Approach 2:
The patent changes the parameters of the prediction model dynamically based on luma values. Multiple linear models with different parameters (slope and intercept) are created to adapt to different luma ranges. By selecting and applying appropriate parameters based on the local luma characteristics, the method achieves better prediction accuracy and compression performance compared to fixed-parameter approaches.
2Loss of energy
If more compression is applied to reduce video data volume, then storage and transmission efficiency improve, but image quality may be compromised
Solution Approach 1:
The patent applies different prediction strategies to different regions of the chroma block based on local luma characteristics. By creating multiple linear models tailored to specific luma value ranges and applying them locally, the method preserves image quality in different regions appropriately. This local quality approach ensures that compression is optimized for each region without uniformly sacrificing quality across the entire block.
3Device complexity
If simple prediction models are used, then encoding complexity is reduced, but prediction accuracy and compression efficiency deteriorate
Solution Approach 1:
The patent introduces dynamic model selection based on luma value thresholds. The system adapts which linear model to use by dynamically comparing luma values against threshold criteria. This dynamic approach allows the relatively simple linear models to achieve better effective complexity by selecting the most appropriate model for each region, improving compression efficiency without requiring overly complex prediction structures.
Data Source
AI summary
A computing device performs a method of decoding video data by receiving bitstream encoding a chroma block, a corresponding luma block, neighboring luma samples, and neighboring chroma samples; decoding the luma block, the plurality of neighboring luma samples, and the plurality of neighboring chroma samples; selecting a group of reference luma samples and a group of reference chroma samples; computing a threshold luma value from the plurality of reconstructed neighboring luma samples, and a threshold chroma value from the plurality of reconstructed neighboring chroma samples; determining a maximum luma value and a minimum luma value from the group of the reference luma samples; generating multi-model linear model (MMLM) including a first linear model between the minimum luma value and the threshold luma value, and a second linear model between the threshold luma value and the maximum luma value; and reconstructing the chroma block from the luma block using MMLM.