3D Entity Mesh Reconstruction With Octree Surface Voxels
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing 3D model reconstruction methods using deep marching tetrahedra (DMTet) generate sharp and slender tetrahedral meshes near the surface, leading to inaccurate reconstructions due to non-smooth shapes and erroneous protrusions, while implicit neural network-based methods are inefficient with high time consumption.
Innovation Solution
A reconstruction method utilizing voxel partitioning based on an octree structure to distribute voxels densely near the surface, maintaining regular cube shapes and ensuring accurate representation of the geometric shape, avoiding 'No Proper' geometry and improving efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If deep marching tetrahedra method is used to represent surface with deformable tetrahedral mesh, then 3D mesh representation is obtained, but sharp and slender tetrahedral meshes are generated near the surface resulting in inaccurate reconstruction
Solution Approach 1:
The space is segmented into a hierarchical octree structure with multiple levels of resolution. Voxels are partitioned adaptively near the surface to capture geometric details while maintaining regular shapes, avoiding the sharp and slender mesh elements produced by tetrahedral subdivision methods.
Solution Approach 2:
Different voxel resolution levels are applied to different regions of the 3D entity model. Higher resolution voxels are concentrated near the surface where geometric accuracy is critical, while coarser voxels are used in interior regions, optimizing both accuracy and computational efficiency.
2Measurement precision
If multiple subdivisions of tetrahedral mesh are performed near the surface, then surface detail is captured, but sharp and slender meshes are generated causing erroneous protrusions
Solution Approach 1:
The octree structure provides hierarchical segmentation where each level divides space into regular octants. This systematic segmentation captures surface details through controlled refinement near the surface without generating the pathological sharp and slender elements that cause erroneous protrusions.
Solution Approach 2:
The hierarchical octree structure embeds finer resolution voxels within coarser voxels in a nested manner. Each parent voxel contains child voxels at the next resolution level, allowing progressive refinement near the surface while maintaining global structural consistency and avoiding local mesh artifacts.
3Manufacturing precision
If implicit neural network-based method is used for 3D reconstruction, then reconstruction is achieved, but high time consumption is incurred
Solution Approach 1:
The patent replaces the implicit neural network-based approach with a deterministic voxel-based geometric processing method. This substitution eliminates the iterative optimization and heavy computational overhead of neural networks while achieving accurate 3D reconstruction through systematic voxel partitioning and mesh extraction.
Solution Approach 2:
The space is segmented into a hierarchical octree structure with multiple levels of resolution. Voxels are partitioned adaptively near the surface to capture geometric details while maintaining regular shapes, avoiding the sharp and slender mesh elements produced by tetrahedral subdivision methods.
Data Source
AI summary
A three-dimensional (3D) entity model reconstruction method performed by a computer device includes obtaining 3D spatial information corresponding to a 3D entity model, performing voxel partitioning on the 3D entity model based on the 3D spatial information to determine voxels corresponding to the 3D entity model and connected in a tree structure, and constructing a 3D mesh corresponding to the 3D entity model based on the voxels. The voxels are distributed on a surface of the 3D entity model and represent a geometric shape of the surface of the 3D entity model.