GPM Image Encoding/Decoding with Template-Matched Partition Refinement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing demand for high-resolution and high-quality images leads to a significant increase in transmission and storage costs due to the rise in the amount of transmitted information, necessitating high-efficient image compression technologies.
Innovation Solution
An image encoding/decoding method utilizing geometric partitioning mode (GPM) refined by template matching (TM) and asymmetric blending areas to generate an intra prediction mode (IPM) list, enhancing encoding/decoding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If high-resolution and high-quality images are transmitted, then image quality is improved, but transmission cost and storage cost increase
Solution Approach 1:
The patent applies template matching to refine partition shapes and blending areas in GPM, dynamically adjusting geometric parameters to achieve better compression efficiency. By optimizing partition angles, distances, and blending region sizes based on template matching results, the patent reduces the bitstream size while maintaining high image quality, thus lowering transmission and storage costs
2Productivity
If GPM is applied to current block, then encoding efficiency is improved, but partition shape and blending area determination becomes complex
Solution Approach 1:
The patent performs template matching in advance to determine optimal partition shapes and blending areas before generating the prediction block. By pre-calculating template costs and selecting the best partition configuration beforehand, the patent simplifies the subsequent encoding process and reduces computational complexity during actual encoding
Solution Approach 2:
The patent uses the image data itself to generate templates for matching, where the template is derived from the current block's boundary samples. This self-referential approach eliminates the need for external reference data and simplifies the determination process by using the block's own characteristics to guide partition optimization
3Measurement precision
If template matching is used to refine partition shape, then prediction accuracy is improved, but computational complexity increases
Solution Approach 1:
The patent applies template matching selectively to refine only the critical parameters (partition angle and distance) rather than optimizing all possible parameters. By focusing computational effort on the most influential geometric parameters, the patent achieves significant prediction accuracy improvement while limiting the increase in computational complexity
4Manufacturing precision
If asymmetric blending areas are used, then encoding precision is improved, but generation complexity of IPM list increases
Solution Approach 1:
The patent explicitly introduces asymmetric blending areas where the left and right blending regions have different sizes. By allowing asymmetric configuration, the patent achieves better encoding precision for geometric partitions with directional characteristics, while managing complexity through systematic generation rules for the IPM list based on the asymmetric structure
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Provided are an image encoding/decoding method, a method for transmitting a bitstream, and a computer-readable recording medium storing a bitstream. The image decoding method according to the present disclosure may comprise the steps of: obtaining geometric partitioning mode (GPM) information on the basis of a GPM being applied to the current block; determining at least one of a partition shape of a blending region of the GPM on the basis of the obtained GPM information; refining at least one of the partition shape or the blending region on the basis of template matching; and generating a prediction block for the current block by performing the GPM on the current block on the basis of the refined partition shape or the refined blending region.