Faithful Contouring: Near-Lossless 3D Voxel Representation Free from Iso-surface

2025-11-06

AI Summary

Existing isosurface-based 3D mesh voxelization methods rely on watertightening or rendering optimizations, compromising geometric fidelity and internal structure preservation. This paper proposes a sparse voxelization representation that bypasses isosurface extraction, implicit field fitting, and remeshing, enabling high-resolution, near-lossless encoding of arbitrary (non-watertight) meshes. Our core innovations are a silhouette-preserving voxel encoding strategy and a dual-mode autoencoder architecture, which jointly ensure sharp feature retention, explicit modeling of interior geometry, and topology-flexible manipulation. Experiments demonstrate that our direct representation achieves signed distance errors on the order of 10⁻⁵; in reconstruction tasks, it reduces Chamfer Distance by 93% and improves F-score by 35% over state-of-the-art methods, significantly advancing fidelity, structural accuracy, and topological expressiveness.

Technology Category

• Machine Learning: Deep Generative Models & Autoencoders
• Computer Vision: 3D Computer Vision
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Virtualization and resource management in Web systems and infrastructures
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

Accurate and efficient voxelized representations of 3D meshes are the foundation of 3D reconstruction and generation. However, existing representations based on iso-surface heavily rely on water-tightening or rendering optimization, which inevitably compromise geometric fidelity. We propose Faithful Contouring, a sparse voxelized representation that supports 2048+ resolutions for arbitrary meshes, requiring neither converting meshes to field functions nor extracting the isosurface during remeshing. It achieves near-lossless fidelity by preserving sharpness and internal structures, even for challenging cases with complex geometry and topology. The proposed method also shows flexibility for texturing, manipulation, and editing. Beyond representation, we design a dual-mode autoencoder for Faithful Contouring, enabling scalable and detail-preserving shape reconstruction. Extensive experiments show that Faithful Contouring surpasses existing methods in accuracy and efficiency for both representation and reconstruction. For direct representation, it achieves distance errors at the $10^{-5}$ level; for mesh reconstruction, it yields a 93% reduction in Chamfer Distance and a 35% improvement in F-score over strong baselines, confirming superior fidelity as a representation for 3D learning tasks.

Problem

• Achieving near-lossless 3D voxel representation without iso-surface extraction
• Preserving geometric fidelity for complex shapes with sharp features
• Enabling high-resolution representation and reconstruction of arbitrary meshes

Innovation

• Sparse voxel representation without iso-surface extraction
• Preserves sharpness and internal structures near-losslessly
• Dual-mode autoencoder enables scalable shape reconstruction