Adaptive CTU Splitting for High-Resolution Image Encoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image encoding/decoding methods are limited by their fixed square Coding Unit (CU) structure, which restricts adaptability to various local content characteristics, leading to inefficiencies in encoding and decoding high-resolution and high-quality images.
Innovation Solution
The proposed method involves splitting a Coding Tree Unit (CTU) into Coding Units (CUs) using various block splitting structures, including quadtree, binary tree, and ternary tree splitting, allowing for more flexible and adaptive encoding/decoding processes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a fixed square Coding Unit (CU) structure is used for image encoding/decoding, then the encoding/decoding process is simple and consistent, but the adaptability to various local content characteristics is limited
Solution Approach 1:
The patent applies segmentation by dividing the image into Coding Tree Units (CTUs) and further segmenting them into Coding Units (CUs) using multiple tree structures (quadtree, binary tree, ternary tree). This hierarchical segmentation allows the system to adapt to different local content characteristics by selecting appropriate splitting patterns for different regions, thereby resolving the contradiction between adaptability and structural simplicity.
Solution Approach 2:
The patent introduces dynamic block splitting structures where the splitting method (quadtree, binary tree, or ternary tree) can be dynamically selected based on local content characteristics. This dynamic approach allows the encoding/decoding system to adapt to varying image features across different regions, improving adaptability while maintaining manageable complexity through standardized splitting algorithms.
2Productivity
If conventional quadtree splitting is used for all blocks, then the encoding/decoding process is consistent and easy to implement, but the efficiency for high-resolution and high-quality images is insufficient
Solution Approach 1:
The patent enhances encoding efficiency by implementing multi-stage segmentation with quadtree, binary tree, and ternary tree structures. This allows more precise adaptation to high-resolution images by creating finer-grained Coding Units where needed, improving compression efficiency and processing performance for complex high-quality images while maintaining systematic implementation through standardized splitting procedures.
Solution Approach 2:
The patent applies local quality by allowing different block splitting structures (quadtree, binary tree, ternary tree) to be applied to different regions of the image based on local content characteristics. This enables high-resolution and high-quality images to be processed with appropriate granularity and complexity in different areas, improving overall encoding efficiency without uniformly increasing system complexity.
3Adaptability or versatility
If multiple block splitting structures (quadtree, binary tree, ternary tree) are used, then the adaptability to diverse local content characteristics is improved, but the computational complexity increases
Solution Approach 1:
The patent manages computational complexity through hierarchical segmentation where the image is first divided into CTUs, then into CUs using multiple tree structures. This organized segmentation approach allows the system to achieve high adaptability to diverse local content characteristics while controlling computational complexity through systematic, rule-based splitting procedures that can be efficiently implemented.
Solution Approach 2:
The patent controls computational complexity by changing parameters such as the maximum depth of each tree structure and the conditions for applying different splitting methods. These parameter adjustments allow the system to achieve high adaptability to diverse content characteristics while maintaining manageable computational complexity through configurable limits and standardized algorithms.
Data Source
AI summary
The present invention relates to an image encoding/decoding method. The image decoding method includes splitting a coding tree unit (CTU) into at least one coding unit (CU) according to a block splitting structure and performing CU-based decoding, in which the block partition structure is configured such that at least one of binary tree splitting and ternary tree splitting is performed after quadtree splitting is performed.