Linear Interpolation Intra-Prediction for Video Encoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing intra-frame prediction method reduces prediction accuracy as the distance from reference samples increases, as it simply duplicates the generated prediction sample value, leading to decreased performance in encoding and decoding high spatial resolution, high frame rate, and high dimensionality video contents.
Innovation Solution
A linear interpolation intra-prediction method is proposed, which derives values for multiple prediction samples using neighboring reference samples based on the intra-prediction mode and generates prediction samples by interpolating these values, applying weights based on the distance between samples to improve accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If prediction sample values are simply duplicated based on intra-prediction mode, then encoding complexity is reduced, but prediction accuracy deteriorates as distance from reference samples increases
Solution Approach 1:
The patent changes the parameter of prediction sample generation from simple duplication to linear interpolation. By introducing interpolation calculations between reference samples and current block samples, the prediction accuracy is improved while maintaining manageable encoding complexity through efficient mathematical operations.
Solution Approach 2:
The patent introduces an intermediary calculation process (linear interpolation) between the reference samples and the prediction samples. This intermediary step allows for more accurate prediction values by combining reference sample information with spatial distance weighting, rather than directly duplicating values.
2Measurement precision
If more reference samples are used for prediction, then prediction accuracy is improved, but processing power requirements increase
Solution Approach 1:
The patent applies local quality by using different weighting factors for different reference samples based on their spatial relationship to the current block. Closer reference samples receive higher weights, while distant ones receive lower weights, allowing accurate prediction without uniformly processing all reference samples with equal computational effort.
3Measurement precision
If linear interpolation is applied to generate prediction samples, then prediction accuracy is improved, but encoding complexity increases
Solution Approach 1:
The patent applies partial action by performing linear interpolation only for specific prediction modes and blocks where it provides significant accuracy improvement. Not all blocks require full interpolation processing, allowing the system to balance accuracy gains with computational complexity by selectively applying the more complex operation.
Data Source
AI summary
An intra prediction mode-based image processing method includes: obtaining, on the basis of the intra prediction mode of a current block, a first prediction sample value and a second prediction sample value by using a reference sample neighboring the current block; and generating a prediction sample for the current block by linear interpolation of the first prediction sample value and the second prediction sample value.