Point Cloud Attribute Compression Using KD-Tree Graph Blocks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing point cloud attribute compression methods face inefficiencies due to high computational complexity and sub-graph issues, which affect compression performance and processing efficiency.
Innovation Solution
A point cloud attribute compression method based on a KD tree and optimized graph transformation, which includes preprocessing, KD tree division, constructing a graph with adaptive transformation blocks, optimizing the kernel parameter, and performing graph transformation to achieve optimal compression performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If octree decomposition and graph transformation are used for point cloud attribute compression, then compression performance is improved, but computational complexity increases
Solution Approach 1:
The patent divides the point cloud into transformation blocks using KD tree decomposition, where each block contains a subset of points. This segmentation allows the graph transformation to be applied independently to each block, reducing the overall computational complexity while maintaining compression performance. The KD tree structure enables efficient organization and processing of points in hierarchical levels.
Solution Approach 2:
The patent introduces adaptive parameters (σ and τ) that can be dynamically adjusted based on the local characteristics of each transformation block. This dynamic adaptation allows the graph transformation to optimize compression performance for different regions of the point cloud without requiring excessive computational resources for uniform high-quality processing throughout the entire dataset.
2Ease of manufacture
If traditional transformation block division is used, then processing is simpler, but sub-graph issues reduce transformation efficiency
Solution Approach 1:
The patent uses KD tree to recursively divide the point cloud into transformation blocks, ensuring that points within each block are spatially coherent. This segmentation strategy prevents sub-graph issues by guaranteeing that connected components in the graph transformation are complete and well-defined, thereby improving transformation efficiency without significantly increasing processing complexity.
Solution Approach 2:
The patent performs preliminary KD tree decomposition and transformation block formation before applying graph transformation. This preliminary organization of points into properly defined blocks ensures that subsequent graph operations can proceed efficiently without encountering sub-graph issues, as the spatial relationships are already established in the block structure.
3Productivity
If fixed transformation blocks are used, then processing is faster, but adaptability to different point cloud distributions is reduced
Solution Approach 1:
The patent implements adaptive transformation blocks where the KD tree decomposition depth and block characteristics can vary according to the local point density and distribution. This dynamic approach allows faster processing in uniform regions while maintaining high adaptability in complex regions, resolving the contradiction between speed and adaptability.
Solution Approach 2:
The patent applies different processing parameters (σ and τ) to different transformation blocks based on their local characteristics. This local quality approach ensures that each block is processed with parameters optimized for its specific point distribution, achieving both efficiency and adaptability across the entire point cloud.
Data Source
AI summary
Provided is a point cloud attribute compression method based on a KD tree and optimized graph transformation, wherein same, with regard to point cloud data, reduces the influence of a sub-graph issue on the graph transformation efficiency by means of a new transformation block division method, optimizes a graph transformation kernel parameter, and improves the compression performance of the graph transformation, and comprises: point cloud pre-processing, point cloud KD tree division, graph construction in the transformation block, graph transformation kernel parameter training, and a point cloud attribute compression process. The present invention optimizes the division method for a point cloud transformation block, and makes the number of points in the transformation block the same, and also realizes that the dimensionality of a transformation matrix is basically the same, so as to facilitate parallel processing of subsequent graph transformations; also optimizes the graph establishment in the transformation block, and avoids the sub-graph issue caused by the existing method; and at the same time optimizes, by training the kernel parameter of the graph transformation, the sparsity of a graph transformation Laplacian matrix, so as to achieve a better point cloud attribute compression performance.