HilbertGen 3D improves 3D model generation with global shape awareness
HilbertGen-3D:Hilbert--Multifractal Conditioning for Topology-Aware 3D Generation
Computer Vision and Pattern Recognition
Summary
Creating realistic 3D shapes is important for games, medical images, and virtual worlds. Current methods mostly look at small parts of the shape and miss the big picture. The authors developed HilbertGen-3D, which uses a special math tool to understand and include the overall shape structure when making new 3D models. This helps produce more accurate and varied 3D objects like chairs, airplanes, and cars.
What this means in practice
- •For game asset creators: Generate more realistic and diverse 3D models for game environments by combining local details and global shape structure.
- •For virtual reality designers: Create immersive VR environments with higher-fidelity 3D shapes reflecting complex global geometries for better user experience.
Authors
Ziyi Zhang, Shixuan Li, Yorie Nakahira, Paul Bogdan, Guannan Qu
Abstract
High-fidelity 3D generative modeling underpins game asset creation, medical imaging, and VR/AR environmental design, where geometric structure directly influences perception and decision-making. Diffusion models represent a powerful paradigm for point cloud generation. However, existing approaches rely heavily on local voxel or patch features and often miss global structural organization, and much less work has studied how global conditioning can be used to improve generation quality. In this paper, we introduce HilbertGen-3D, a novel topology-aware conditional diffusion framework based on Hilbert-Multifractal Conditioning (HMC). HMC computes multi-scale voxel measures, serializes them via a 3D Hilbert curve to expose structured spatial dependencies, and estimates the generalized fractal dimensions from scale-dependent partition functions. We encode these signals as global embeddings and multi-scale tokens, and inject them into a diffusion denoiser through an efficient bottleneck fusion module. Across the Chair, Airplane, and Car categories of ShapeNet, our two conditioning variants reduce the absolute deviation of 1-Nearest Neighbor Accuracy (1-NNA) from 50% by up to 8.90 percentage points and increase coverage by up to 5.88 percentage points relative to TopoDiT-3D. This shows that combining local density cues with global Hilbert-multifractal structure provides a simple and effective recipe for high-fidelity and diverse 3D generation.