Score-Based Point Cloud Denoising

TL;DR

This paper introduces a neural network-based score estimation method for denoising point clouds by iteratively increasing the likelihood of clean points, outperforming existing techniques across various noise models.

Contribution

It proposes a novel score-based neural network architecture and training framework for effective point cloud denoising without requiring explicit noise models.

Findings

Outperforms state-of-the-art denoising methods

Effective across multiple noise models

Potential application in point cloud upsampling

Abstract

Point clouds acquired from scanning devices are often perturbed by noise, which affects downstream tasks such as surface reconstruction and analysis. The distribution of a noisy point cloud can be viewed as the distribution of a set of noise-free samples $p(x)$ convolved with some noise model $n$, leading to $(p * n)(x)$ whose mode is the underlying clean surface. To denoise a noisy point cloud, we propose to increase the log-likelihood of each point from $p * n$ via gradient ascent -- iteratively updating each point's position. Since $p * n$ is unknown at test-time, and we only need the score (i.e., the gradient of the log-probability function) to perform gradient ascent, we propose a neural network architecture to estimate the score of $p * n$ given only noisy point clouds as input. We derive objective functions for training the network and develop a denoising algorithm leveraging on the estimated scores. Experiments demonstrate that the proposed model outperforms state-of-the-art methods under a variety of noise models, and shows the potential to be applied in other tasks such as point cloud upsampling. The code is available at \url{https://github.com/luost26/score-denoise}.