Adaptive N-Ary Tree Leaves for Efficient 3D Point Cloud Encoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing three-dimensional data encoding methods face challenges in improving coding efficiency and reducing processing amounts, particularly in the context of point cloud data compression.
Innovation Solution
A three-dimensional data encoding method that appends information to a bitstream indicating whether a leaf in an N-ary tree structure includes a single or multiple points, generating and encoding the appropriate tree structure accordingly, and a decoding method that decodes this information to improve coding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a fixed N-ary tree structure with single-point leaves is used, then the encoding structure is simple, but coding efficiency deteriorates when multiple points exist in a leaf region
Solution Approach 1:
The patent applies dynamics by making the leaf node structure adaptive rather than fixed. Each leaf node can dynamically adjust between containing a single point or multiple points based on the actual data distribution, allowing the encoding structure to flexibly respond to varying point densities in different spatial regions
Solution Approach 2:
The patent implements local quality by allowing different parts of the N-ary tree to have different leaf configurations. Regions with high point density can have multi-point leaves while sparse regions use single-point leaves, optimizing compression efficiency locally rather than applying a uniform structure globally
2Productivity
If a fixed N-ary tree structure with multi-point leaves is used, then coding efficiency improves for dense regions, but processing amount increases and complexity rises
Solution Approach 1:
The patent applies segmentation by dividing the point cloud data into different spatial regions and applying appropriate leaf configurations to each segment. The encoding process is segmented into phases: first encoding the flag indicating leaf type, then conditionally encoding either single-point or multi-point structures based on the flag value
Solution Approach 2:
The patent uses parameter changes by introducing a flag parameter that controls the leaf node configuration. This parameter can be switched between different states (single-point vs. multi-point) based on data characteristics, allowing the system to adapt its processing mode without fundamental structural changes
3Quantity of substance
If traditional point cloud compression is applied, then data accumulation and transmission are enabled, but processing amount and computational complexity increase
Solution Approach 1:
The patent applies partial action by selectively applying complex multi-point leaf encoding only where necessary (in regions with multiple points), while using simpler single-point encoding in other regions. This avoids the excessive processing required by uniform multi-point structures applied throughout the entire point cloud
Data Source
AI summary
A three-dimensional data encoding method includes: appending, to a bitstream, first information indicating whether a leaf to be included in an N-ary tree structure of three-dimensional points included in three-dimensional data is to include a single three-dimensional point or two or more three-dimensional points, where N is an integer greater than or equal to 2; when the first information indicates that the leaf is to include a single three-dimensional point, generating an N-ary tree structure in which a leaf includes a single three-dimensional point, and encoding the N-ary tree structure; and when the first information indicates that the leaf is to include two or more three-dimensional points, generating an N-ary tree structure in which a leaf includes two or more three-dimensional points, and encoding the N-ary tree structure.