Intra Prediction Blending Control Using Mode Amplitude Ratios
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The setting of weighted blending conditions in existing intra prediction methods is too broad, leading to reduced prediction quality and introduction of noise when weighted blending is performed on image contents that do not require it, affecting the accuracy of intra prediction.
Innovation Solution
Determine a weighted blending condition based on the amplitude values of the first and second intra prediction modes to decide whether to perform weighted blending prediction for the current block, ensuring accurate intra prediction by avoiding unnecessary blending.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If weighted blending prediction is performed on all current blocks, then prediction accuracy may be improved for some blocks, but prediction quality deteriorates and noise is introduced for blocks that do not require blending
Solution Approach 1:
The patent applies parameter changes by modifying the weighted blending condition from a broad, fixed criterion to a refined criterion based on amplitude value comparisons. Specifically, it calculates amplitude values for different intra prediction modes and uses their ratios to dynamically determine whether weighted blending should be applied. This parameter-based approach allows the system to adapt the blending operation to the actual content characteristics of each block, improving prediction accuracy for suitable blocks while avoiding quality degradation for unsuitable blocks.
2Productivity
If weighted blending condition is set broadly to enable more blocks to use weighted blending, then prediction accuracy may improve for more blocks, but unnecessary blending is performed on blocks that do not require it, reducing overall prediction quality
Solution Approach 1:
The patent implements local quality by applying different prediction strategies to different blocks based on their local characteristics. Instead of uniformly applying weighted blending to all blocks or using a single broad condition, the system evaluates the amplitude values of intra prediction modes for each block individually. The weighted blending operation is then selectively applied only to blocks where the amplitude value ratio indicates suitability, ensuring that each block receives the appropriate prediction treatment for its local content characteristics.
3Reliability
If weighted blending is avoided for all blocks to maintain prediction quality, then noise is prevented, but prediction accuracy is lost for blocks that would benefit from weighted blending
Solution Approach 1:
The patent applies dynamics by making the weighted blending decision dynamic rather than static. Instead of using a fixed rule that either always enables or always disables weighted blending, the system dynamically evaluates the amplitude values of different intra prediction modes for each block and adjusts the blending decision accordingly. The amplitude value ratio serves as a dynamic parameter that adapts to the content characteristics of each block, allowing the system to optimize between prediction quality and prediction accuracy on a block-by-block basis.
Data Source
AI summary
An intra prediction method includes the following. A first intra prediction mode and a second intra prediction mode for the current block are determined according to amplitude values of N intra prediction modes. A weighted blending condition for the current block is determined according to an amplitude value of the first intra prediction mode and an amplitude value of the second intra prediction mode. A target prediction value of the current block is determined according to the weighted blending condition and at least one of the first intra prediction mode, the second intra prediction mode, or a third intra prediction mode.