Neural Wavelet-domain Diffusion for 3D Shape Generation

TL;DR

This paper introduces a novel wavelet-domain diffusion method for 3D shape generation that produces diverse, high-quality shapes with complex topology and fine details, outperforming existing models.

Contribution

It proposes a compact wavelet-based implicit representation and a diffusion-based generator with a detail predictor for enhanced 3D shape synthesis.

Findings

Outperforms state-of-the-art models in shape quality and diversity

Generates shapes with complex topology and fine details

Produces high-quality, clean surface reconstructions

Abstract

This paper presents a new approach for 3D shape generation, enabling direct generative modeling on a continuous implicit representation in wavelet domain. Specifically, we propose a compact wavelet representation with a pair of coarse and detail coefficient volumes to implicitly represent 3D shapes via truncated signed distance functions and multi-scale biorthogonal wavelets, and formulate a pair of neural networks: a generator based on the diffusion model to produce diverse shapes in the form of coarse coefficient volumes; and a detail predictor to further produce compatible detail coefficient volumes for enriching the generated shapes with fine structures and details. Both quantitative and qualitative experimental results manifest the superiority of our approach in generating diverse and high-quality shapes with complex topology and structures, clean surfaces, and fine details, exceeding the 3D generation capabilities of the state-of-the-art models.