Differentiable Manifold Reconstruction for Point Cloud Denoising

TL;DR

This paper introduces a differentiable manifold reconstruction method for denoising 3D point clouds, explicitly recovering the underlying surface structure to improve denoising quality over existing approaches.

Contribution

It proposes a novel autoencoder-based neural network that learns the manifold of noisy point clouds using differentiable pooling and local features, enabling explicit surface reconstruction.

Findings

Outperforms state-of-the-art denoising methods on synthetic and real noise.

Effectively captures intrinsic surface structures in noisy point clouds.

Supports both supervised and unsupervised training modes.

Abstract

3D point clouds are often perturbed by noise due to the inherent limitation of acquisition equipments, which obstructs downstream tasks such as surface reconstruction, rendering and so on. Previous works mostly infer the displacement of noisy points from the underlying surface, which however are not designated to recover the surface explicitly and may lead to sub-optimal denoising results. To this end, we propose to learn the underlying manifold of a noisy point cloud from differentiably subsampled points with trivial noise perturbation and their embedded neighborhood feature, aiming to capture intrinsic structures in point clouds. Specifically, we present an autoencoder-like neural network. The encoder learns both local and non-local feature representations of each point, and then samples points with low noise via an adaptive differentiable pooling operation. Afterwards, the decoder infers the underlying manifold by transforming each sampled point along with the embedded feature of its neighborhood to a local surface centered around the point. By resampling on the reconstructed manifold, we obtain a denoised point cloud. Further, we design an unsupervised training loss, so that our network can be trained in either an unsupervised or supervised fashion. Experiments show that our method significantly outperforms state-of-the-art denoising methods under both synthetic noise and real world noise. The code and data are available at https://github.com/luost26/DMRDenoise