Taming Anomalies with Down-Up Sampling Networks: Group Center Preserving Reconstruction for 3D Anomaly Detection

2025-07-05

AI Summary

High-precision point clouds suffer from geometric distortion during reconstruction, severely limiting detection accuracy in 3D anomaly detection. To address this, we propose the Down-Up Sampling Network (DUS-Net), which introduces a novel group-center preservation mechanism to explicitly enforce local geometric consistency between down-sampling and up-sampling stages. A noise generation module is integrated to enhance training diversity, while a multi-scale feature fusion Up-Net ensures high-fidelity point cloud reconstruction. Evaluated on Real3D-AD and Anomaly-ShapeNet, DUS-Net achieves state-of-the-art performance: 79.9%/79.5% object-level AUROC and 71.2%/84.7% point-level AUROC, respectively—surpassing all existing methods. The core contributions lie in (i) the group-center preservation strategy for geometric structure retention, (ii) the noise-augmented training paradigm for robustness, and (iii) the hierarchical Up-Net architecture enabling precise, detail-preserving reconstruction.

Technology Category

• Computer Vision: 3D Computer Vision
• Data Mining & Knowledge Management: Anomaly/Outlier Detection
• Machine Learning: Deep Generative Models & Autoencoders

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets

Abstract

Reconstruction-based methods have demonstrated very promising results for 3D anomaly detection. However, these methods face great challenges in handling high-precision point clouds due to the large scale and complex structure. In this study, a Down-Up Sampling Network (DUS-Net) is proposed to reconstruct high-precision point clouds for 3D anomaly detection by preserving the group center geometric structure. The DUS-Net first introduces a Noise Generation module to generate noisy patches, which facilitates the diversity of training data and strengthens the feature representation for reconstruction. Then, a Down-sampling Network~(Down-Net) is developed to learn an anomaly-free center point cloud from patches with noise injection. Subsequently, an Up-sampling Network (Up-Net) is designed to reconstruct high-precision point clouds by fusing multi-scale up-sampling features. Our method leverages group centers for construction, enabling the preservation of geometric structure and providing a more precise point cloud. Extensive experiments demonstrate the effectiveness of our proposed method, achieving state-of-the-art (SOTA) performance with an Object-level AUROC of 79.9% and 79.5%, and a Point-level AUROC of 71.2% and 84.7% on the Real3D-AD and Anomaly-ShapeNet datasets, respectively.

Problem

• Reconstruct high-precision 3D point clouds for anomaly detection
• Preserve geometric structure via group center reconstruction
• Handle large-scale complex point clouds with noise injection

Innovation

• Down-Up Sampling Network for 3D reconstruction
• Noise Generation module enhances feature diversity
• Group center preservation improves geometric accuracy