Point Cloud Coding with Selective Isolated-Point Signaling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing point cloud compression technologies are inefficient and costly due to the large data volumes and time-consuming storage and transmission requirements, particularly in applications requiring real-time communication and dynamic mapping for autonomous driving and cultural heritage.
Innovation Solution
The proposed method leverages existing video codecs to compress geometry, occupancy, and texture of a dynamic point cloud as separate video sequences, with metadata encoded efficiently and the bulk information handled by video codecs, using techniques like HEVC and VVC for efficient compression.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If point cloud data is stored and transmitted using existing compression technologies, then data can be captured and represented in 3D space, but the data volume is large and storage and transmission are costly and time-consuming
Solution Approach 1:
The point cloud data is segmented into multiple octants using an octree structure, where each node represents a specific 3D region. This segmentation allows for region-specific encoding strategies, enabling more efficient compression by treating different spatial regions differently based on their characteristics and importance.
Solution Approach 2:
The patent applies different coding modes (occupancy coding, direct coding, inferred direct coding) to different regions of the point cloud based on local characteristics. Isolated points in certain regions are encoded with higher precision using direct coding, while other regions use more compact occupancy coding, optimizing the balance between quality and compression ratio.
2Measurement precision
If isolated points are coded using direct coding mode in all nodes, then coding accuracy is improved, but the complexity increases and not all nodes require such detailed coding
Solution Approach 1:
The patent applies direct coding mode selectively only to isolated points rather than all nodes. The system identifies isolated points (nodes with specific occupancy patterns) and applies high-precision direct coding only to those locations, while using simpler occupancy coding for non-isolated regions, thus achieving high accuracy where needed without unnecessary complexity elsewhere.
Solution Approach 2:
Instead of applying full direct coding to all nodes (excessive action), the patent applies direct coding only partially to isolated points where it is most beneficial. This partial application reduces overall complexity while maintaining sufficient coding accuracy for the critical isolated point regions.
3Adaptability or versatility
If a dedicated syntax is proposed to signal the use of Direct Coding Mode for all nodes, then coding flexibility is improved, but the bitstream overhead increases
Solution Approach 1:
The patent signals direct coding mode information locally at isolated point locations rather than globally for all nodes. This selective signaling approach provides the necessary coding flexibility to handle isolated points appropriately while minimizing bitstream overhead by only transmitting mode information where it is actually needed.
Solution Approach 2:
The patent extracts and handles isolated point coding separately from the general occupancy coding process. By identifying isolated points and applying dedicated coding syntax only to those specific locations, the system achieves coding flexibility for isolated points without imposing significant overhead on the entire bitstream.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Aspects of the disclosure provide methods and apparatuses for point cloud compression and decompression. In some examples, an apparatus for point cloud compression/decompression includes processing circuitry. In some examples, the processing circuitry receives a bitstream carrying compressed data for a point cloud. The processing circuitry determines that a current node in an octree structure is eligible for an isolated mode. The octree structure corresponds to three dimensional (3D) partitions of a space of the point cloud. Then the processing circuitry determines, based on information of one or more other nodes, a single isolated point flag for the current node that indicates whether the current node is coded with a single isolated point.