Point Cloud Prediction Units for Lower-Latency Transmission
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies face challenges in efficiently processing and transmitting large volumes of point cloud data, particularly for applications like virtual reality, augmented reality, and self-driving services, due to the high latency and complexity of encoding and decoding such data, especially when captured by LiDAR equipment.
Innovation Solution
The method involves encoding geometry data by splitting it into prediction units and applying motion vectors selectively, along with attribute data encoding and signaling, to optimize point cloud data transmission and reception, using techniques like inter-prediction and global motion vectors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If point cloud data is transmitted with high fidelity for VR/AR/self-driving applications, then quality is improved, but data volume increases causing transmission inefficiency
Solution Approach 1:
The point cloud data is segmented into multiple prediction units (PUs) of varying sizes. Each PU is independently processed with appropriate prediction modes, allowing efficient compression while maintaining local quality. The segmentation enables selective application of compression techniques to different regions based on their importance and complexity.
Solution Approach 2:
The patent employs multiple prediction modes (global motion vector, partitioned motion vector, intra prediction) with different parameter complexities. The encoder selects and signals the appropriate mode for each prediction unit, changing the processing parameters adaptively to balance compression efficiency and reconstruction quality for different data regions.
2Ease of manufacture
If conventional point cloud encoding methods are used, then implementation is simple, but latency and encoding complexity increase for large datasets
Solution Approach 1:
By dividing the point cloud into prediction units and processing them independently with parallelizable operations, the encoding pipeline can be optimized for throughput. The segmented structure allows multiple PUs to be processed simultaneously, reducing overall encoding latency while maintaining implementation feasibility.
Solution Approach 2:
The patent performs preliminary organization of point cloud data into prediction units and pre-calculates motion vectors for reference frames. This preliminary processing prepares the data in an optimized structure that accelerates the main encoding process, reducing latency without complicating the core encoding logic.
3Productivity
If motion vectors are applied to all prediction units, then compression performance improves, but encoding complexity increases
Solution Approach 1:
The patent applies motion vector prediction selectively based on local characteristics of each prediction unit. Global motion vectors are used for regions with uniform motion, while partitioned motion vectors are applied only to regions requiring finer motion representation. This local adaptation optimizes compression performance while avoiding unnecessary complexity in uniform regions.
Solution Approach 2:
Instead of applying complex partitioned motion vectors to all prediction units, the encoder uses a hierarchical approach: global motion vectors are applied universally, and partitioned motion vectors are applied only when necessary (partial action). This selective application achieves sufficient compression performance without the excessive complexity of universal fine-grained prediction.
Data Source
AI summary
Disclosed herein is a method of transmitting point cloud data. The method may include encoding geometry data of the point cloud data, encoding attribute data of the point cloud data based on the geometry data, and transmitting the encoded geometry data, the encoded attribute data and signaling data, the geometry encoding includes splitting the geometry data into one or more prediction units, and inter-prediction encoding the geometry data by selectively applying a motion vector to each of the split prediction units, and the signaling data includes information for identifying whether the motion vector is applied for each prediction unit.