Point Cloud LOD Partitioning Using Morton Code Neighbor Search
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In Geometry-based Point Cloud Compression (G-PCC) encoding, the computational complexity of Level of Detail (LOD) partitioning is high due to inaccurate neighbor node searches, leading to increased prediction residuals and encoding bits, reducing encoding efficiency.
Innovation Solution
A partitioning method that calculates Morton codes for points in a point cloud, determines right shift numbers for LOD layers, and uses these codes to search for neighbor nodes, reducing the need for spatial distance calculations and improving prediction accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional spatial distance-based neighbor node search is used in LOD partitioning, then prediction accuracy may be maintained, but computational complexity increases and encoding efficiency decreases
Solution Approach 1:
The patent transforms the neighbor node search from spatial distance-based parameters to Morton code-based parameters. By encoding point cloud coordinates into Morton codes and using code distance instead of spatial distance, the search process becomes computationally simpler while maintaining partitioning effectiveness. This parameter transformation resolves the contradiction by reducing computational complexity without sacrificing encoding efficiency.
Solution Approach 2:
The patent replaces the mechanical spatial distance calculation system with a computational Morton code comparison system. Instead of calculating Euclidean distances between points in 3D space, the system uses bitwise operations and code comparisons on Morton codes, which are computationally less intensive and can be implemented more efficiently in software, thereby reducing overall computational complexity.
2Measurement precision
If incomplete neighbor node search is performed, then computational complexity is reduced, but prediction accuracy deteriorates and encoding bits increase
Solution Approach 1:
The patent performs preliminary sorting of point cloud data by Morton codes before the LOD partitioning process. This preliminary action organizes the data in a way that enables efficient neighbor node identification during partitioning. By pre-organizing the data structure, the system can perform more complete and accurate neighbor node searches without excessive computational cost, thereby improving prediction accuracy while maintaining encoding efficiency.
Solution Approach 2:
The patent introduces Morton codes as an intermediary representation between spatial coordinates and neighbor node relationships. The Morton code acts as a mediator that preserves spatial locality information in a compressed, computationally friendly format. This intermediary enables accurate neighbor node identification through simple code comparisons rather than complex spatial calculations, resolving the contradiction between prediction accuracy and encoding efficiency.
3Device complexity
If Morton code-based partitioning is used, then computational complexity is reduced, but implementation complexity increases
Solution Approach 1:
The patent segments the point cloud processing into distinct stages: coordinate to Morton code conversion, Morton code-based LOD partitioning, and attribute prediction. Each stage handles a specific aspect of the processing, making the overall complex task more manageable. The segmentation allows each component to be optimized independently and facilitates implementation by breaking down the complex Morton code-based partitioning into smaller, more manageable operations.
Data Source
AI summary
Disclosed are a partitioning method, an encoder, a decoder and a computer storage medium. The method includes: determining location information of a point of a point cloud to be partitioned; when i is less than or equal to M−1, determining right-shift number Ni of ith LOD layer in the point cloud, M representing a preset maximum quantity of layers for LOD partitioning; for the ith LOD layer, shifting location information of the point rightwards by Ni-digit, performing storing in a preset storage area based on right-shifted location information; determining location information of a parent point corresponding to a current point in the ith LOD layer; according to determined location information of the parent point, searching the preset storage area for a neighbor point of the parent point; partitioning the current point into an (i+1)th LOD layer, or the neighbor point into the ith LOD layer.