3D Point Cloud Decoding With Adaptive Triangle Extension
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing lossy compression methods for 3D point clouds, such as MPEG-I part 5 (VPCC) and MPEG-I part 9 (G-PCC), struggle to maintain acceptable visual quality while achieving low bitrates necessary for real-time transmission of dynamic AR/VR point clouds, especially for dense and complex geometries, leading to issues with reconstruction accuracy.
Innovation Solution
The method involves decoding a bitstream containing octree information and vertex data to reconstruct triangles, followed by voxelization, with an adaptive halo parameter based on the sampling distance to extend triangles when necessary, ensuring accurate reconstruction of the point cloud geometry.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If lossy compression is used to achieve low bitrates for real-time transmission, then transmission bandwidth is reduced, but reconstruction quality of the point cloud deteriorates
Solution Approach 1:
The patent applies preliminary action by extending triangles along their sides before the voxelization process. This pre-extension ensures that when voxelization occurs, the extended triangles already cover the areas where points might be missed due to sampling distance, thereby preventing reconstruction quality deterioration before it happens
Solution Approach 2:
The patent changes the geometric parameters of triangles by extending them along their sides based on sampling distance. This parameter modification allows the triangles to cover a larger area, ensuring that points at various sampling distances are captured during voxelization, thus maintaining reconstruction quality at low bitrates
2Device complexity
If traditional voxelization is used without triangle extension, then processing complexity is low, but points are missed during reconstruction leading to quality loss
Solution Approach 1:
The patent performs triangle extension as a preliminary step before voxelization. By extending triangles along their sides based on sampling distance before the actual voxelization process, the method ensures that all potential points are covered in advance, eliminating missed points without adding significant processing complexity during the main reconstruction phase
Data Source
AI summary
A method for decoding a 3-dimensional (3D) point cloud from a bitstream is performed by a decoder. The method includes: receiving and decoding the bitstream, wherein the bitstream contains octree information including information about an octree structure of a volume of the point cloud and vertex information including information about vertex presence and a position of a vertex on edges of cuboids of leaf nodes of the octree structure; determining triangles by connecting vertices of one cuboid relating to a leaf node of the octree structure; determining points of the point cloud by voxelization of the triangles; determining whether additional information contained in the bitstream meets a pre-defined condition; and-when the pre-defined condition is met, extending at least one triangle along at least one side for voxelization based on the sampling distance.