Pruned Sparse Voxel Octree for 3D Surface Compression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current technologies face challenges in efficiently representing and compressing time-varying 3D surfaces with arbitrary topology and scale, particularly in virtual reality and augmented reality applications, where compact representations and random access capabilities are essential for complex and dynamic environments.
Innovation Solution
The implementation of a block-oriented representation using a pruned Sparse Voxel Octree (SVO) combined with explicit or implicit surface representations, such as TriSoup and Bezier Volumes, allows for multiresolution surface compression and decompression, enabling efficient storage, transmission, and rendering of surfaces with scalable resolution and parallel processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If a mesh is broken into regions for compression, then spatial random access is achieved, but device complexity increases
Solution Approach 1:
The mesh is divided into multiple regions or patches that can be independently compressed and accessed. Each region can be processed separately, enabling spatial random access where specific regions can be decoded without processing the entire mesh, thus improving ease of operation while managing complexity through modular organization.
Solution Approach 2:
The patent implements a hierarchical structure where regions are organized in nested levels of detail. Coarser representations are embedded within finer representations, allowing progressive refinement and random access to different levels of detail without requiring the entire high-resolution mesh to be processed.
2Measurement precision
If progressive mesh compression is achieved by subdividing faces or splitting vertices, then resolution is improved, but loss of information increases
Solution Approach 1:
The patent pre-computes and stores a hierarchy of mesh representations at different levels of detail before compression. This preliminary action creates a progressive structure where coarse approximations are prepared in advance, allowing the decoder to start with low-resolution geometry and progressively refine it without losing essential geometric information.
Solution Approach 2:
The patent transforms the mesh representation by changing parameters such as vertex positions and face connections across different levels of detail. By systematically varying these parameters from coarse to fine representations, the patent achieves progressive resolution improvement while maintaining geometric fidelity through controlled parameter transitions.
3Reliability
If the finest mesh output has connectivity identical to the input mesh, then reliability is improved, but device complexity increases
Solution Approach 1:
The patent creates simplified copies of the original mesh connectivity at different levels of detail rather than attempting to preserve exact connectivity throughout the compression process. These copied connectivity structures are progressively refined to match the original, maintaining reliability of the final output while reducing algorithmic complexity through the use of approximate intermediate representations.
Data Source
AI summary
A machine can be specially configured to generate, compress, decompress, store, access, communicate, or otherwise process a special data structure that represents a three-dimensional surface of an object. The data structure can be or include a pruned sparse voxel octree in which each node in the octree corresponds to a different block of the octree, and children of the node in the octree correspond to the smaller blocks that subdivide the block. Moreover, each block occupied by the surface or a portion thereof can define its enclosed surface or portion thereof explicitly or implicitly.