Point Cloud Data Transmission Using Adaptive LOD Octrees
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for processing point cloud data face challenges in efficiently transmitting and receiving large amounts of data, addressing latency and encoding/decoding complexity, improving compression performance, and managing duplicated points generated during geometry quantization and voxelization in geometry-based point cloud compression (G-PCC).
Innovation Solution
A method and apparatus for transmitting and receiving point cloud data that involves encoding geometry information through voxelization, partitioning voxels into sub-voxels, generating occupancy bits, and reconstructing geometry, while encoding attribute information using levels of detail (LODs) and predicting attribute values, with signaling information managing duplicated point processing options.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If point cloud data is transmitted with high compression ratio, then transmission efficiency is improved, but decoding complexity and latency increase
Solution Approach 1:
The point cloud data is segmented into multiple octrees at different levels of detail (LOD). Each octree represents a different level of geometric detail, allowing the receiver to selectively decode only the necessary levels based on computational resources and latency requirements, thereby reducing overall decoding complexity while maintaining transmission efficiency
Solution Approach 2:
Multiple levels of detail (LODs) are pre-computed and encoded during the transmission phase. The receiver can then directly use these pre-prepared LODs without performing complex real-time computations, reducing decoding latency and complexity while maintaining high compression ratios
2Productivity
If geometry quantization is applied to reduce data size, then compression performance is improved, but geometry loss becomes unpredictable
Solution Approach 1:
The patent implements adaptive geometry quantization where the quantization step size dynamically adjusts based on the local geometric complexity and importance of different point cloud regions. Critical geometric features use finer quantization to minimize geometry loss, while less important regions use coarser quantization to maintain compression performance
Solution Approach 2:
Different quantization parameters are applied to different regions of the point cloud based on their geometric importance. Regions containing critical geometric features (such as object boundaries or high-curvature areas) receive higher precision quantization, while homogeneous regions use lower precision, making geometry loss predictable and controllable
3Quantity of substance
If duplicated points are removed during voxelization, then data redundancy is reduced, but processing complexity increases
Solution Approach 1:
The patent segments the deduplication process into multiple passes corresponding to different LODs. Points are first grouped into voxels at a coarse level, then progressively refined at finer levels. This segmented approach reduces the computational burden compared to processing all points simultaneously, while effectively removing duplicates
4Adaptability or versatility
If multiple levels of detail (LODs) are generated, then decoding flexibility is improved, but encoding complexity increases
Solution Approach 1:
The encoding process is segmented into hierarchical levels where each LOD is encoded independently using octree structures. This segmentation allows the encoder to process and transmit multiple levels of detail in a structured manner, managing encoding complexity through systematic organization while providing decoding flexibility
Solution Approach 2:
Multiple LODs are nested within a hierarchical octree structure where coarser levels are contained within finer levels. This nested organization allows efficient memory management and processing during encoding, as each level builds upon the previous level, reducing overall encoding complexity while maintaining adaptability
Data Source
AI summary
A point cloud data transmission method according to embodiments may comprise the steps of: acquiring point cloud data; encoding geometry information of the point cloud data; encoding attribute information of the point cloud data on the basis of the geometry information; and transmitting a bitstream including the encoded geometry information, the encoded attribute information, and signaling information.