Intra Prediction Using Linear Parameters for Image Encoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image compression technologies face challenges in efficiently encoding and decoding high-resolution and high-quality images, particularly in configuring intra prediction-based prediction blocks and signaling intra prediction methods effectively.
Innovation Solution
An image decoding method and device that derive a linear prediction parameter from a neighboring region to obtain a prediction sample for the current block, using weights for horizontal and vertical variations and a predetermined offset, and adaptively signaling the intra prediction method based on flags from the bitstream.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional intra prediction technology is used, then encoding complexity is reduced, but prediction accuracy and image quality deteriorate
Solution Approach 1:
The patent changes the parameters of intra prediction by introducing linear prediction parameters (weights and offsets) that adapt to local image characteristics. Instead of using fixed prediction modes, the system calculates optimal linear prediction parameters based on neighboring block variations, thereby improving prediction accuracy while maintaining manageable encoding complexity through parameter adaptation.
Solution Approach 2:
The patent implements dynamic intra prediction by calculating linear prediction parameters adaptively for different blocks based on their local characteristics. The weights and offsets are dynamically determined from neighboring block variations, allowing the prediction method to adapt to different regions of the image, thus improving overall prediction accuracy without requiring exhaustive search.
2Measurement precision
If linear prediction parameters are derived from neighboring regions, then intra prediction performance improves, but encoding complexity increases
Solution Approach 1:
The patent applies local quality by deriving linear prediction parameters specifically from neighboring regions that are most relevant to the current block. By focusing on local variations in adjacent blocks and using localized windows for parameter calculation, the method improves prediction performance for each block while avoiding the need to process the entire image, thus controlling encoding complexity.
Solution Approach 2:
The patent segments the neighboring region into specific windows (e.g., top, left, top-left neighbors) and derives prediction parameters from appropriate segments based on the current block's position and characteristics. This segmentation allows selective use of neighboring information, improving prediction accuracy while reducing the amount of data that needs to be processed compared to using all neighboring pixels.
3Productivity
If high-resolution images are compressed, then transmission efficiency improves, but prediction accuracy deteriorates
Solution Approach 1:
The patent performs preliminary action by deriving linear prediction parameters from already-decoded neighboring blocks before encoding the current block. This allows the system to utilize available high-resolution neighboring information to guide the prediction process, improving prediction accuracy for high-resolution images while maintaining transmission efficiency through effective compression of the residual.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An image decoding/encoding method and apparatus according to the present disclosure may: derive linear prediction parameters for the current block from neighboring regions adjacent to the current block; and generate a prediction sample of the current block on the basis of the linear prediction parameters. Here, the linear prediction parameters may include at least one of a weighted value relating to the amount of change in sample values of the neighboring regions and a certain offset.