Intra Prediction Linear Interpolation for Coding Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
High-resolution and high-quality images require efficient compression techniques to reduce data transmission and storage costs, as existing methods are inefficient in managing the increased information bits.
Innovation Solution
An image coding method and apparatus that improves intra-prediction performance by using linear interpolation of neighboring reference samples, including left, upper-left, upper, right, lower-right, and lower neighboring samples, to generate prediction and residual samples, thereby reducing additional data and enhancing coding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If high-resolution and high-quality images are transmitted or stored using conventional methods, then image quality is maintained, but transmission cost and storage cost increase due to increased data amount
Solution Approach 1:
The patent changes the parameter of prediction sample generation from conventional methods to linear interpolation-based methods. By using linear interpolation of reference samples to generate prediction samples, the invention achieves better prediction accuracy, which reduces the residual data amount while maintaining high image quality reconstruction
Solution Approach 2:
The patent uses reference samples (copies of neighboring pixel data) to predict and reconstruct the current block samples. By copying and interpolating reference sample values to generate prediction samples, the system reduces the amount of original data needed while maintaining reconstruction quality
2Device complexity
If conventional intra-prediction methods are used, then coding process is simple, but prediction performance is insufficient leading to larger residual data
Solution Approach 1:
The patent modifies the prediction method parameters by implementing linear interpolation calculations between reference samples. This changes the mathematical operation from simple copying or averaging to interpolation, improving prediction accuracy and reducing information loss in the residual data
Solution Approach 2:
The patent replaces simple mechanical copying operations with mathematical interpolation operations. By substituting the mechanical approach of direct sample copying with mathematical linear interpolation, the system achieves superior prediction performance while maintaining computational efficiency
3Loss of information
If more neighboring reference samples are used for linear interpolation, then intra-prediction performance improves, but data amount of additional information increases
Solution Approach 1:
The patent uses exactly six specific neighboring reference samples (left, upper-left, upper, right, lower-right, and lower) which is a carefully selected partial set. This partial action approach provides sufficient prediction accuracy improvement without requiring all possible reference samples, thus avoiding excessive data transmission
Solution Approach 2:
The six neighboring reference samples serve multiple functions simultaneously: they provide prediction data for linear interpolation, defines prediction directions, and enable reconstruction of various block types. This multi-functionality reduces the need for additional dedicated data structures
Data Source
AI summary
The intra prediction method according to the present invention comprises the steps of: deriving an intra prediction mode of a current block; deriving neighboring reference samples of the current block; and generating a prediction sample of the current block on the basis of linear interpolation of a first neighboring reference sample positioned in the prediction direction of the intra prediction mode, and a second neighboring reference sample positioned in the opposite direction of the prediction direction, among the neighboring reference samples of the current block, wherein the neighboring reference samples include left neighboring reference samples, left upper neighboring reference samples, upper neighboring reference samples, right neighboring reference samples, right lower neighboring reference samples, and lower neighboring reference samples of the current block. The present invention can improve intra prediction performance, and thereby reduce the amount of data required for residual information and increase the overall coding efficiency.