Point-Cloud Decoding Order Using Mask_Planar Without Sorting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional point-cloud decoding technologies require complete decoding of position information before attribute information can be decoded, and lack a defined processing order without sorting, leading to inefficiencies.
Innovation Solution
A point-cloud decoding device and method that enable unique association between positional and attribute information by generating Mask_planar and flags for child nodes during quadtree or binary tree division, allowing decoding without prior sorting.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If sorting is performed before decoding attribute information, then unique association between position and attribute information is ensured, but decoding time increases due to the additional sorting process
Solution Approach 1:
The patent applies preliminary action by pre-defining the decoding order of attribute information based on the spatial arrangement of child nodes in the tree structure. The Mask_planar values are set in advance to encode the decoding sequence (e.g., 00, 01, 10, 11 corresponding to different child node pairs), allowing the decoder to process attributes without performing explicit sorting operations while maintaining unique association between positions and attributes.
2Productivity
If sorting is omitted to reduce decoding time, then processing efficiency improves, but the processing order of attribute information becomes undefined
Solution Approach 1:
The patent changes the parameter representation by using Mask_planar values (0 or 1) for each child node to encode the decoding sequence. Instead of sorting points based on spatial coordinates, the system transforms the spatial relationship into a coded parameter sequence that directly defines the attribute decoding order, thereby maintaining reliability without requiring sorting operations.
3Manufacturing precision
If complete position information decoding is required before attribute decoding, then accurate point-cloud reconstruction is achieved, but overall processing time increases
Solution Approach 1:
The patent segments the decoding process into independent parallel streams: position information decoding and attribute information decoding. By using the tree structure with Mask_planar to define attribute decoding order, the system allows attribute decoding to proceed independently without waiting for complete position decoding, thereby reducing overall processing time while maintaining reconstruction accuracy through the defined spatial-attribute correspondence.
Data Source
AI summary
A point-cloud decoding device according to the present invention includes: a tree synthesizing unit configured to generate Mask_planar corresponding to eight child nodes of a node, wherein in a case where the node is subjected to quadtree division or binary tree division, the tree synthesizing unit (2020) is configured to set, as “0”, a value of the Mask_planar corresponding to a child node belonging to a plane closer to an origin in an axial direction in which no division is made, and set, as “1”, a value of the Mask_planar corresponding to a child node belonging to a plane further away from the origin in the axial direction in which no division is made.