Node-Based Point Cloud Geometry and Attribute Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current point cloud compression technologies face challenges in reducing data size for efficient transmission and storage, particularly in achieving low latency encoding and decoding, as they often require completing geometry coding before attribute coding, which can lead to increased latency and inefficiency.
Innovation Solution
The proposed method involves node-based geometry and attribute coding, where geometry information is encoded first for each node at a specified depth, followed by attribute encoding, allowing for partial bitstream transmission and decoding without waiting for complete geometry encoding of the entire point cloud, using techniques like octree, predictive tree, and RAHT coding.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If geometry coding is completed before attribute coding, then encoding order is maintained, but latency increases and transmission efficiency decreases
Solution Approach 1:
The patent segments the point cloud encoding process into independent node-level units, where each node's geometry and attribute coding can be performed independently. This segmentation allows parallel processing of multiple nodes simultaneously, eliminating the sequential dependency that caused latency. The encoder divides the point cloud into discrete nodes and processes them in parallel streams, enabling early transmission of completed node data without waiting for entire geometry or attribute encoding to finish.
Solution Approach 2:
The patent performs preliminary geometry coding for individual nodes before complete attribute coding is finished for the entire point cloud. This preliminary action on a per-node basis allows encoded geometry data to be transmitted and decoded early, reducing latency. The system prepares and transmits node-level geometry information in advance, enabling progressive reconstruction without waiting for all attributes to be encoded.
2Speed
If complete point cloud partitioning is performed before encoding, then data structure is established, but transmission speed decreases due to waiting for full encoding
Solution Approach 1:
The patent applies partial action by encoding and transmitting only a subset of nodes (those that have been processed) rather than waiting for complete encoding of all nodes. The encoder transmits bitstreams for decoded nodes as soon as they are ready, providing partial but useful data to the decoder. This approach prioritizes transmission speed over complete data availability, allowing progressive reconstruction of the point cloud while maintaining data reliability for transmitted portions.
3Productivity
If neighboring nodes' information is considered during encoding, then compression efficiency improves, but encoding complexity and latency increase
Solution Approach 1:
The patent segments the encoding process into independent node units with optional neighbor consideration. Each node can be encoded with or without reference to neighboring nodes based on configuration. This segmentation allows flexible trade-offs: when compression efficiency is prioritized, neighbor information is used locally; when speed is prioritized, independent node encoding is used. The segmentation enables parallel processing that mitigates the time penalty of neighbor consideration.
Data Source
AI summary
A method and apparatus for coding information of a point cloud may be performed by at least one processor and comprises: obtaining the point cloud including a set of points in a three-dimensional space; partitioning the point cloud into a tree structure comprising a plurality of nodes at different depths; encoding geometry information of the nodes; and encoding attribute information of the nodes before the entire point cloud is partitioned.