Cross-Component Linear Intra Prediction With Segmented LUT Indexing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies face challenges in achieving efficient compression of video data without sacrificing picture quality, particularly in the context of limited network bandwidth and memory resources.
Innovation Solution
The method employs a cross-component linear model for intra prediction, utilizing a lookup table (LUT) to determine linear model parameters α and β based on reconstructed luma samples, minimizing the size of the LUT and enhancing the efficiency of fetching values, thereby improving chroma sample prediction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a lookup table is used to store linear model parameters for all possible luma sample differences, then the precision of chroma prediction is improved, but the memory size and complexity of the system increases significantly
Solution Approach 1:
The patent segments the lookup table into multiple smaller tables, each handling a specific range of luma sample differences. Instead of one large table covering all possible differences, multiple smaller tables divide the range, reducing the size of each individual table while maintaining comprehensive coverage for precise predictions across all difference values.
Solution Approach 2:
The patent pre-calculates and stores linear model parameters in the lookup table during system initialization or encoding phase. By preparing these parameters in advance based on historical or statistical data, the system avoids real-time complex calculations during decoding, improving prediction precision while keeping the stored table size manageable through the segmentation principle.
2Measurement precision
If more bits are extracted from the binary representation of luma sample difference for LUT indexing, then the accuracy of parameter selection is improved, but the complexity of the indexing mechanism increases
Solution Approach 1:
The patent applies different numbers of extracted bits for indexing different lookup tables based on their specific requirements. Each table can use an optimized number of bits from the binary representation of luma sample difference, allowing high accuracy where needed while reducing complexity for tables with coarser requirements. This local optimization balances precision and complexity across the segmented table structure.
Data Source
AI summary
Apparatuses and methods for encoding and decoding are provided. The method for intra predicting a chroma sample of a block by applying cross-component linear model includes: obtaining reconstructed luma samples; determining maximum and minimum luma sample values based on the reconstructed luma samples; obtaining a difference of the maximum and minimum luma sample values. The method also includes: fetching a value out of a lookup table (LUT) by using a set of bits as an index, the set of bits following a position of the most-significant bit; obtaining linear model parameters based on the fetched value; and calculating a predicted chroma sample value by using the obtained linear model parameters. The efficiency to fetch the value out of the LUT is increased.