Patch-Wise Point Cloud Generation: A Divide-and-Conquer Approach

TL;DR

This paper introduces a novel divide-and-conquer framework for high-fidelity 3D point cloud generation, utilizing patch-wise transformers and learnable priors to better capture local and global geometric structures, outperforming existing methods.

Contribution

It proposes a new patch-wise point cloud generation method with transformer-based interactions, advancing understanding of 3D shape geometry and improving generation quality.

Findings

Outperforms state-of-the-art methods on ShapeNet dataset

Effectively captures local and global geometric structures

Demonstrates high-fidelity point cloud generation

Abstract

A generative model for high-fidelity point clouds is of great importance in synthesizing 3d environments for applications such as autonomous driving and robotics. Despite the recent success of deep generative models for 2d images, it is non-trivial to generate 3d point clouds without a comprehensive understanding of both local and global geometric structures. In this paper, we devise a new 3d point cloud generation framework using a divide-and-conquer approach, where the whole generation process can be divided into a set of patch-wise generation tasks. Specifically, all patch generators are based on learnable priors, which aim to capture the information of geometry primitives. We introduce point- and patch-wise transformers to enable the interactions between points and patches. Therefore, the proposed divide-and-conquer approach contributes to a new understanding of point cloud generation from the geometry constitution of 3d shapes. Experimental results on a variety of object categories from the most popular point cloud dataset, ShapeNet, show the effectiveness of the proposed patch-wise point cloud generation, where it clearly outperforms recent state-of-the-art methods for high-fidelity point cloud generation.