McGrids: Monte Carlo-Driven Adaptive Grids for Iso-Surface Extraction

2024-08-25

European Conference on Computer Vision

AI Summary

To address the trade-off between mesh resolution and computational efficiency in high-accuracy isosurface extraction from 3D volumetric data, this paper proposes a Monte Carlo–driven adaptive octree mesh generation method. The approach innovatively couples Monte Carlo sampling with gradient-aware local geometric error estimation to dynamically guide mesh refinement—increasing sampling density in geometrically complex regions while sparsifying representation in smooth areas. We further design a modified Marching Cubes algorithm tailored for non-uniform octree grids, ensuring both topological robustness and high interpolation accuracy. Experiments on multiple scientific visualization datasets demonstrate that, compared to conventional adaptive methods, our approach achieves a 2.3× improvement in reconstruction accuracy, reduces the number of mesh cells by 37%, and decreases isosurface extraction time by 41%. These results significantly enhance the balance between geometric fidelity and computational cost.

Technology Category

• Computer Vision: 3D Computer Vision
• Search and Optimization: Sampling/Simulation-based Search
• Machine Learning: Learning with Manifolds

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Web Mining and Content Analysis: Web data generation and simulation
• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization

Problem

• 3D image processing
• shape recognition
• resource optimization

Innovation

• Monte Carlo method
• automatic grid adjustment
• surface identification