Video Intra Prediction Using Adaptive Reference Samples
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video signal encoding/decoding methods face challenges in efficiently performing intra prediction, particularly in determining a suitable range of reference samples for calculating DC values, and in adapting weights for these samples.
Innovation Solution
The method involves determining an intra prediction mode for a current block, calculating a DC value using adjacent reference samples, and deriving prediction samples based on these DC values. The range of reference samples is adaptively determined for each region, and weights are applied differently to each sample to improve prediction accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a fixed range of reference samples is used for DC value calculation, then the encoding/decoding process is simple, but the prediction accuracy is insufficient for high-resolution images
Solution Approach 1:
The patent applies dynamics by making the reference sample range adaptive rather than fixed. The range of reference samples used for DC value calculation is dynamically adjusted based on the intra prediction mode and block characteristics, allowing the system to optimize prediction accuracy for different scenarios while maintaining simplicity for standard cases
Solution Approach 2:
The patent changes parameters by varying the weight coefficients applied to different reference samples based on their positions and the prediction mode. This parameter adaptation allows the system to improve prediction accuracy by emphasizing more relevant reference samples while reducing the impact of less relevant ones
2Measurement precision
If equal weights are applied to all reference samples, then the calculation is simple, but the prediction accuracy is reduced
Solution Approach 1:
The patent applies local quality by assigning different weight coefficients to different reference samples based on their local characteristics and positions. Reference samples closer to the current block or more relevant to the prediction direction receive higher weights, while others receive lower weights, creating a localized optimization strategy
Solution Approach 2:
The patent changes parameters by dynamically adjusting weight coefficients based on the intra prediction mode and reference sample positions. This parameter variation allows the system to adapt to different prediction scenarios and improve accuracy without requiring complex machine learning models
3Measurement precision
If DC value is calculated for the entire block, then the process is simple, but the prediction accuracy for different regions is insufficient
Solution Approach 1:
The patent applies segmentation by dividing the current block into multiple regions and calculating separate DC values for each region. This allows different parts of the block to have region-specific prediction characteristics, improving accuracy for blocks with varying content patterns while maintaining computational efficiency through systematic region划分
Solution Approach 2:
The patent applies local quality by calculating region-specific DC values that reflect the local characteristics of different parts of the block. Each region's DC value is derived from locally relevant reference samples, allowing the prediction to adapt to local variations in image content
Data Source
AI summary
A method of video decoding according to the present invention may comprise determining an intra prediction mode of a current block, determining a DC value using reference samples adjacent to the current block when the intra prediction mode of the current block is a DC mode, and deriving a prediction sample of the current block based on the DC value.