Geometric Partitioning of Image Blocks for Hybrid Intra-Inter Prediction
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 that applies geometric partitioning to blocks, determining whether to resample reference samples, and performs intra/inter prediction based on geometric partitioning, deriving intra prediction modes from collocated blocks and reference picture resampling.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If high-resolution and high-quality images are transmitted and stored, then image quality is improved, but transmission cost and storage cost increase
Solution Approach 1:
The current block is divided into a first sub-block and a second sub-block through geometric partitioning. The first sub-block uses intra prediction while the second sub-block uses inter prediction, allowing different compression strategies to be applied to different regions, thereby reducing overall bitstream size while maintaining image quality.
Solution Approach 2:
Different prediction modes are applied to different sub-blocks based on local characteristics. The first sub-block uses intra prediction mode derived from collocated blocks, while the second sub-block uses inter prediction, optimizing compression efficiency for each region's specific content characteristics.
2Productivity
If geometric partitioning with intra/inter prediction is applied, then encoding/decoding efficiency is improved, but device complexity increases
Solution Approach 1:
The intra prediction mode for the first sub-block is derived in advance from the collocated block's intra prediction mode. This preliminary derivation simplifies the encoding process by reusing existing prediction mode information rather than calculating new modes from scratch.
Solution Approach 2:
The intra prediction mode from the collocated block is copied and applied to the first sub-block. This copying approach reduces computational complexity by reusing prediction mode information from spatially or temporally correlated blocks.
Data Source
AI summary
An image encoding/decoding method, a method of transmitting a bitstream and a computer-readable recording medium storing a bitstream are provided. The image decoding method performed by an image decoding apparatus may comprise determining whether to apply geometric partitioning to a current block, partitioning the current block into a first sub-block and a second sub-block based on geometric partitioning being applied to the current block, and performing intra prediction on the first sub-block and performing inter prediction on the second sub-block. An intra prediction mode of the first sub-block may be derived based on an intra prediction mode of a collocated block within a reference block and whether to resample the reference picture.