Point Cloud Attribute Compression Using KD-Tree Graph Transforms
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing point cloud attribute compression methods face challenges in achieving high compression performance while maintaining low operation complexity and avoiding subgraph problems, particularly in utilizing spatial correlation effectively.
Innovation Solution
A point cloud attribute compression method based on enhanced graph transformation, which involves airspace division using a KD tree, spectral clustering, and local graph transformation with two transformation modes, optimizing compression performance by selecting the best mode based on transformation efficiency.
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 operation complexity increases and subgraph problems occur
Solution Approach 1:
The patent divides the point cloud into multiple transformation blocks using KD tree decomposition, where each block is processed independently. This segmentation approach reduces operation complexity by limiting the graph transformation scope to smaller local regions rather than processing the entire point cloud at once, while still achieving good compression performance through localized spatial correlation exploitation.
Solution Approach 2:
The patent applies different graph transformation strategies to different transformation blocks based on local characteristics. By performing KD tree decomposition and processing each block independently with locally optimized graph transformations, the method adapts to local spatial correlations without being constrained by global subgraph problems, thus balancing compression performance and operational complexity.
2Loss of information
If KD tree decomposition with graph transformation is used, then compression performance is greatly improved and subgraph problem is solved, but operation complexity remains high
Solution Approach 1:
The patent uses KD tree decomposition to divide the point cloud into multiple transformation blocks at a certain depth, ensuring each block contains approximately the same number of points. This segmentation reduces the computational burden of graph transformation by processing smaller independent blocks rather than the entire point cloud, thereby lowering operation complexity while maintaining compression performance.
Solution Approach 2:
The patent performs graph transformation only on attribute information within each transformation block rather than the entire point cloud. By applying graph transformation selectively to local regions with similar spatial correlations, the method achieves effective compression without the excessive computational cost of global graph transformation, thus reducing operation complexity.
3Loss of information
If spatial decomposition is performed to utilize spatial correlation, then compression performance can be improved, but the method may not fully exploit spatial correlation between points
Solution Approach 1:
The patent dynamically adjusts the graph construction and transformation process based on local spatial correlations within each transformation block. By using KD tree decomposition to create blocks with similar spatial characteristics and then performing graph transformation adapted to each block's specific geometry, the method fully exploits spatial correlation between points, improving compression performance compared to fixed approaches.
Solution Approach 2:
The patent changes the spatial decomposition parameters by using KD tree with a specific depth to control block size and point distribution. This parameter adjustment ensures that each transformation block contains points with strong spatial correlations, allowing the graph transformation to effectively exploit these correlations and achieve better compression performance.
Data Source
AI summary
An enhanced graph transformation-based point cloud attribute compression method. For point cloud attribute information, a point cloud is first subjected to airspace division by using a K-dimension (KD) tree; a new graph transformation processing method in combination with spectral analysis is provided; the point cloud is then subjected to spectral clustering on graphs in coded blocks of the point cloud; expansion is performed on the basis of existing graph transformation to implement a local graph transformation scheme; enhanced graph transformation with two transformation modes is formed; the compression performance of graph transformation is improved. The method comprises: performing color space transformation of point cloud attributes; dividing the point cloud by using the KD tree to obtain the coded blocks; performing spectral clustering-based enhanced graph transformation; performing transformation mode decision; and performing uniform quantization and entropy coding. Provided is a new spectral analysis-based enhanced graph transformation scheme, wherein two transformation modes are comprised, and the optimal mode is selected by the mode decision; after the point cloud is divided with the tree, a graph is created in each coded block and the graph transformation is used as transformation mode I; on this basis, graph spectral clustering is implemented; the graph is divided into two local graphs and then local graph transformation is performed to serve as transformation mode II; in the enhanced graph transformation scheme supporting the two transformation modes, the optimal mode is selected by the mode decision to achieve the optimal performance of point cloud attribute compression.