UGD: An Unsupervised Geometric Distance for Evaluating Real-world Noisy Point Cloud Denoising

TL;DR

This paper introduces UGD, an unsupervised geometric distance metric for evaluating noisy point cloud denoising without needing ground-truth data, using learned priors from clean point clouds.

Contribution

It proposes a novel unsupervised evaluation method based on a learned GMM prior and a self-supervised feature extraction network, enabling real-world noisy data assessment.

Findings

UGD achieves comparable results to supervised metrics on synthetic noise data.

UGD effectively evaluates denoising quality on real-world noisy point clouds.

The method operates solely on noisy data, removing the need for ground-truth clean point clouds.

Abstract

Point cloud denoising is a fundamental and crucial challenge in real-world point cloud applications. Existing quantitative evaluation metrics for point cloud denoising methods are implemented in a supervised manner, which requires both the denoised point cloud and the corresponding ground-truth clean point cloud to compute a representative geometric distance. This requirement is highly problematic in real-world scenarios, where ground-truth clean point clouds are often unavailable. In this paper, we propose a simple yet effective unsupervised geometric distance (UGD) for real-world noisy point cloud denoising, calculated solely from noisy point clouds. The core idea of UGD is to learn a patch-wise prior model from a set of clean point clouds and then employ this prior model as the ground-truth to quantify the degradation by measuring the geometric variations of the denoised point cloud. To this end, we first learn a pristine Gaussian Mixture Model (GMM) with extracted patch-wise quality-aware features from a set of pristine clean point clouds by a patch-wise feature extraction network, which serves as the ground-truth for the quantitative evaluation. Then, the UGD is defined as the weighted sum of distances between each patch of the denoised point cloud and the learned pristine GMM model in the patch space. To train the employed patch-wise feature extraction network, we propose a self-supervised training framework through multi-task learning, which includes pair-wise quality ranking, distortion classification, and distortion distribution prediction. Quantitative experiments with synthetic noise confirm that the proposed UGD achieves comparable performance to supervised full-reference metrics. Moreover, experimental results on real-world data demonstrate that the proposed UGD enables unsupervised evaluation of point cloud denoising methods based exclusively on noisy point clouds.