Point Cloud Encoding Using Scaling Offsets for Spinning Sensors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing point cloud compression technologies, such as G-PCC, face challenges in efficiently encoding and decoding sparse geometry data captured by spinning sensors heads, particularly in terms of latency, simplicity, and compression performance, especially when dealing with sparse point clouds and attributes, due to the complexity of finding predictors and the need for buffering all coordinates before attribute decoding.
Innovation Solution
The method involves signaling scaling offset information in the bitstream to allow for immediate decoding and scaling of spherical coordinates, enabling attribute encoding and decoding without requiring complete coordinate buffering, thus reducing latency and memory footprint while maintaining decoding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of substance
If G-PCC compression technology is used for point cloud data, then compression performance is improved, but latency increases due to the need for buffering all coordinates before attribute decoding
Solution Approach 1:
The patent segments the point cloud data processing into separate geometry decoding and attribute decoding stages. By using spherical coordinate representation with scaling offsets, the geometry can be decoded independently and immediately, while attribute decoding can proceed in parallel or subsequently without waiting for complete coordinate buffering, thus reducing overall latency while maintaining compression efficiency.
Solution Approach 2:
The patent applies preliminary scaling offset calculations to the spherical coordinates before attribute decoding. By pre-computing and signaling the scaling offsets in the bitstream, the decoder can immediately scale coordinates as they are decoded without requiring buffer accumulation, enabling faster processing and reduced latency while preserving compression performance.
2Loss of substance
If G-PCC compression technology is used for point cloud data, then compression performance is improved, but device complexity increases due to the complexity of finding predictors and buffering requirements
Solution Approach 1:
The patent extracts and separates the scaling offset information from the main coordinate data flow. By signaling scaling offsets independently in the bitstream and applying them directly to spherical coordinates during decoding, the complex predictor-finding process is simplified and buffered coordinate storage is eliminated, reducing device complexity while maintaining compression efficiency.
Solution Approach 2:
The patent changes the coordinate system parameters from Cartesian to spherical coordinates and introduces scaling offset parameters. This parameter transformation simplifies the mathematical relationships in point cloud data captured by spinning sensors, reducing the complexity of predictor calculation and buffering operations while preserving high compression ratios.
3Measurement precision
If complete coordinate buffering is performed before attribute decoding, then decoding accuracy is improved, but memory footprint increases
Solution Approach 1:
The patent implements a dynamic decoding approach where spherical coordinates are decoded and scaled on-demand rather than buffering all coordinates statically. The scaling offset mechanism allows coordinates to be processed incrementally as needed for attribute decoding, maintaining numerical precision while dynamically managing memory usage to reduce footprint.
Data Source
AI summary
Methods and apparatus of encoding/decoding a point cloud into a bitstream of encoded point cloud data captured by a spinning sensors head. Each point of the point cloud is associated with spherical coordinates and attributes. The methods comprises signaling, in the bitstream, a scaling offset information representing scaling offsets, and for each current point of the point cloud: encoding/decoding the spherical coordinates of the current point; obtaining decoded spherical coordinates of the current point from the encoded spherical coordinates; scaling the decoded spherical coordinates of the current point by using the scaling offsets; and encoding/decoding at least one attribute of the current point based on the scaled decoded spherical coordinates.