Unsupervised Inference of Signed Distance Functions from Single Sparse Point Clouds without Learning Priors

TL;DR

This paper introduces a novel neural network approach that infers signed distance functions directly from single sparse point clouds without prior learning, improving generalization and accuracy in surface reconstruction tasks.

Contribution

It proposes an end-to-end method that leverages parameterized surfaces and thin plate splines to infer SDFs without supervision or learned priors, especially effective for sparse point clouds.

Findings

Outperforms state-of-the-art methods in surface reconstruction accuracy.

Enhances generalization to unseen sparse point clouds.

Works effectively on both synthetic datasets and real scans.

Abstract

It is vital to infer signed distance functions (SDFs) from 3D point clouds. The latest methods rely on generalizing the priors learned from large scale supervision. However, the learned priors do not generalize well to various geometric variations that are unseen during training, especially for extremely sparse point clouds. To resolve this issue, we present a neural network to directly infer SDFs from single sparse point clouds without using signed distance supervision, learned priors or even normals. Our insight here is to learn surface parameterization and SDFs inference in an end-to-end manner. To make up the sparsity, we leverage parameterized surfaces as a coarse surface sampler to provide many coarse surface estimations in training iterations, according to which we mine supervision and our thin plate splines (TPS) based network infers SDFs as smooth functions in a statistical way. Our method significantly improves the generalization ability and accuracy in unseen point clouds. Our experimental results show our advantages over the state-of-the-art methods in surface reconstruction for sparse point clouds under synthetic datasets and real scans.The code is available at \url{https://github.com/chenchao15/NeuralTPS}.