Fast Point Cloud Generation with Straight Flows

TL;DR

This paper introduces Point Straight Flow (PSF), a novel 3D point cloud generation model that achieves high-quality results in a single step by reformulating diffusion processes into a straight path, enabling real-world applications with low latency.

Contribution

The paper proposes a reformulation of diffusion models into a straight path and a distillation strategy to generate high-quality point clouds in one step, reducing latency significantly.

Findings

PSF performs comparably to standard diffusion models.

Outperforms other efficient 3D point cloud methods.

Effective in real-world tasks like point cloud completion.

Abstract

Diffusion models have emerged as a powerful tool for point cloud generation. A key component that drives the impressive performance for generating high-quality samples from noise is iteratively denoise for thousands of steps. While beneficial, the complexity of learning steps has limited its applications to many 3D real-world. To address this limitation, we propose Point Straight Flow (PSF), a model that exhibits impressive performance using one step. Our idea is based on the reformulation of the standard diffusion model, which optimizes the curvy learning trajectory into a straight path. Further, we develop a distillation strategy to shorten the straight path into one step without a performance loss, enabling applications to 3D real-world with latency constraints. We perform evaluations on multiple 3D tasks and find that our PSF performs comparably to the standard diffusion model, outperforming other efficient 3D point cloud generation methods. On real-world applications such as point cloud completion and training-free text-guided generation in a low-latency setup, PSF performs favorably.