Point Cloud Coding Using Largest Coding Units
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Point clouds, which are used to represent 3D environments, often contain large amounts of data, making them costly and time-consuming to store and transmit.
Innovation Solution
The proposed solution involves methods and apparatuses for point cloud compression and decompression, specifically using node-based geometry and attribute coding. This includes performing geometry coding on a point cloud at a first partition depth, determining largest coding units (LCUs) at a second partition depth, setting a coding state for each LCU, and performing geometry coding based on this state. Additionally, the density of each LCU is determined to decide between octree-based and predictive tree-based geometry coding modes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If point cloud data is stored and transmitted in high quality, then representation accuracy is improved, but data storage cost and transmission time increase
Solution Approach 1:
The point cloud is divided into multiple Largest Coding Units (LCUs) at different partition depths. Each LCU is processed independently with its own coding state, allowing parallel processing and reducing overall transmission time while maintaining quality through selective refinement of important regions.
Solution Approach 2:
Different coding modes (octree-based vs. predictive tree-based) are applied to different LCUs based on their local density characteristics. High-density regions use octree-based coding for better precision, while low-density regions use predictive tree-based coding to reduce data量, achieving local optimization of the representation accuracy to data size ratio.
2Measurement precision
If point cloud data is stored and transmitted in high quality, then representation accuracy is improved, but data storage cost increases
Solution Approach 1:
Different coding modes (octree-based vs. predictive tree-based) are applied to different LCUs based on their local density characteristics. High-density regions use octree-based coding for better precision, while low-density regions use predictive tree-based coding to reduce data量, achieving local optimization of the representation accuracy to data size ratio.
Solution Approach 2:
The coding state parameters (including context for entropy coding and geometry occupancy history information) are dynamically adjusted based on LCU density and position. This allows the system to adapt the compression parameters locally, maintaining representation accuracy in critical areas while reducing storage requirements in less critical areas.
3Measurement precision
If geometry coding is performed at multiple partition depths, then coding precision is improved, but processing complexity increases
Solution Approach 1:
The point cloud is divided into multiple Largest Coding Units (LCUs) at different partition depths. Each LCU is processed independently with its own coding state, allowing parallel processing and reducing overall transmission time while maintaining quality through selective refinement of important regions.
Solution Approach 2:
Different coding modes (octree-based vs. predictive tree-based) are applied to different LCUs based on their local density characteristics. High-density regions use octree-based coding for better precision, while low-density regions use predictive tree-based coding to reduce data量, achieving local optimization of the representation accuracy to data size ratio.
Data Source
AI summary
According to an aspect of the disclosure, a method of point cloud geometry encoding in a point cloud encoder is provided. In the method, geometry coding is performed on a point cloud at a first partition depth. Further, a plurality of largest coding units (LCUs) of the point cloud is determined at a second partition depth. A coding state of a LCU of the plurality of LCUs of the point cloud is set at the second partition depth. The geometry coding is performed on the plurality of LCUs of the point cloud at the second partition depth based on the coding state of the LCU at the second partition depth.