U-CAN: Unsupervised Point Cloud Denoising with Consistency-Aware Noise2Noise Matching

2025-10-29

AI Summary

Point cloud denoising typically relies on supervised learning with paired noisy-clean data, yet acquiring ground-truth noise-free scans from real-world acquisitions remains infeasible. To address this, we propose U-CAN, an unsupervised framework that achieves effective denoising without clean point cloud supervision. Our method introduces two key innovations: (1) a consistency-aware Noise2Noise matching mechanism, which jointly models multi-step denoising trajectories and enforces statistical alignment across multiple noisy observations to enhance robustness in noise modeling; and (2) a geometric consistency constraint that preserves local structural fidelity and enables natural transfer to 2D image denoising. Extensive experiments demonstrate that U-CAN significantly outperforms existing unsupervised methods on point cloud denoising and upsampling benchmarks, while also achieving performance on par with state-of-the-art supervised approaches—further validating its efficacy on standard image denoising benchmarks.

Technology Category

• Machine Learning: Unsupervised & Self-Supervised Learning
• Computer Vision: Diffusion Models for Vision
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Graph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

Point clouds captured by scanning sensors are often perturbed by noise, which have a highly negative impact on downstream tasks (e.g. surface reconstruction and shape understanding). Previous works mostly focus on training neural networks with noisy-clean point cloud pairs for learning denoising priors, which requires extensively manual efforts. In this work, we introduce U-CAN, an Unsupervised framework for point cloud denoising with Consistency-Aware Noise2Noise matching. Specifically, we leverage a neural network to infer a multi-step denoising path for each point of a shape or scene with a noise to noise matching scheme. We achieve this by a novel loss which enables statistical reasoning on multiple noisy point cloud observations. We further introduce a novel constraint on the denoised geometry consistency for learning consistency-aware denoising patterns. We justify that the proposed constraint is a general term which is not limited to 3D domain and can also contribute to the area of 2D image denoising. Our evaluations under the widely used benchmarks in point cloud denoising, upsampling and image denoising show significant improvement over the state-of-the-art unsupervised methods, where U-CAN also produces comparable results with the supervised methods.

Problem

• Unsupervised denoising of noisy point clouds
• Eliminating manual clean data requirements
• Enhancing geometry consistency across domains

Innovation

• Unsupervised framework with Noise2Noise matching
• Multi-step denoising path via neural network
• Consistency-aware constraint for geometry denoising