Intra-Prediction Using Sample Groups to Cut HD/UHD Complexity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image encoding and decoding methods face inefficiencies and high complexity, particularly in intra-prediction processes for high-definition and ultra-high-definition images, which require improved compression techniques.
Innovation Solution
A multi-sample prediction method is introduced, where a sample group is determined within a block, a representative position is identified, and a representative prediction value is calculated for the entire group, reducing the need for individual sample calculations and interpolation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional intra-prediction methods are used for HD and UHD images, then prediction accuracy is maintained, but encoding complexity and processing time increase significantly
Solution Approach 1:
The current block is divided into multiple sample groups, where each sample group contains multiple samples. Instead of processing each sample individually, the method processes each sample group as a unit, determining a representative prediction value for the entire group. This segmentation approach reduces the total number of prediction operations required while maintaining prediction accuracy across the block.
Solution Approach 2:
Multiple samples within a sample group are merged by applying the same representative prediction value to all samples in the group. This combining approach consolidates what would otherwise require multiple separate prediction calculations into a single operation per sample group, significantly reducing computational complexity for large blocks in HD and UHD images.
2Measurement precision
If individual sample prediction is performed for each sample in a block, then prediction precision is maximized, but the number of calculations and processing time increase
Solution Approach 1:
The method applies partial action by performing detailed prediction operations only on representative samples or sample groups rather than every single sample. The representative prediction value derived from partial processing is then extended to cover the entire sample group, achieving sufficient prediction precision without the excessive computational effort of processing every sample individually.
Data Source
AI summary
The multi sample prediction method of the present invention comprises the steps of: determining a sample group consisting of a plurality of samples inside a decoding target block; determining a representative position corresponding to the sample group, inside the decoding target block; determining a representative prediction value for the sample group, on the basis of the determined representative position; and determining the determined representative prediction value as the final prediction value for each of the plurality of samples making up the sample group. The present invention enhances efficiency in encoding/decoding and reduces complexity thereof.