FrePolad: Frequency-Rectified Point Latent Diffusion for Point Cloud Generation

TL;DR

FrePolad introduces a novel frequency-rectified latent diffusion approach for point cloud generation, combining a VAE and DDPM to achieve high quality, diversity, and efficiency with variable point counts.

Contribution

It presents a new frequency rectification method and a latent diffusion model that enhance point cloud generation quality, diversity, and scalability.

Findings

State-of-the-art quality and diversity in point cloud generation.

High computational efficiency and scalability.

Effective handling of variable point cloud cardinality.

Abstract

We propose FrePolad: frequency-rectified point latent diffusion, a point cloud generation pipeline integrating a variational autoencoder (VAE) with a denoising diffusion probabilistic model (DDPM) for the latent distribution. FrePolad simultaneously achieves high quality, diversity, and flexibility in point cloud cardinality for generation tasks while maintaining high computational efficiency. The improvement in generation quality and diversity is achieved through (1) a novel frequency rectification via spherical harmonics designed to retain high-frequency content while learning the point cloud distribution; and (2) a latent DDPM to learn the regularized yet complex latent distribution. In addition, FrePolad supports variable point cloud cardinality by formulating the sampling of points as conditional distributions over a latent shape distribution. Finally, the low-dimensional latent space encoded by the VAE contributes to FrePolad's fast and scalable sampling. Our quantitative and qualitative results demonstrate FrePolad's state-of-the-art performance in terms of quality, diversity, and computational efficiency. Project page: https://chenliang-zhou.github.io/FrePolad/.