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

TL;DR

This paper introduces DUS-Net, a novel network for 3D anomaly detection that reconstructs high-precision point clouds by preserving geometric group centers, improving detection accuracy on challenging datasets.

Contribution

The paper proposes DUS-Net with a Noise Generation module, Down-Net, and Up-Net to enhance 3D anomaly detection by maintaining geometric structure during reconstruction.

Findings

Achieves state-of-the-art AUROC scores on Real3D-AD and Anomaly-ShapeNet datasets.

Effectively preserves geometric structure for high-precision point cloud reconstruction.

Demonstrates significant improvement over existing methods in 3D anomaly detection.

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.