Bridging 3D Anomaly Localization and Repair via High-Quality Continuous Geometric Representation

2025-05-30

AI Summary

To address inaccurate anomaly localization in 3D point cloud anomaly detection under pose variations and complex geometric anomalies, this paper proposes Pose-Aware SDF—a novel framework that employs a continuous Signed Distance Function (SDF) for pose-invariant, high-fidelity implicit shape modeling, jointly enabling anomaly localization and in-situ geometric reconstruction. The method comprises three core components: a pose alignment module, an implicit SDF network, and an anomaly-aware scoring module—enabling, for the first time, pixel-accurate localization and continuous geometric repair. Evaluated on Real3D-AD and Anomaly-ShapeNet, it achieves object-level AUROC scores of 80.2% and 90.0%, respectively, substantially outperforming state-of-the-art methods. The implementation is publicly released to advance research in 3D anomaly detection.

Technology Category

• Computer Vision: 3D Computer Vision
• Data Mining & Knowledge Management: Anomaly/Outlier Detection
• Machine Learning: Calibration & Uncertainty Quantification

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

3D point cloud anomaly detection is essential for robust vision systems but is challenged by pose variations and complex geometric anomalies. Existing patch-based methods often suffer from geometric fidelity issues due to discrete voxelization or projection-based representations, limiting fine-grained anomaly localization. We introduce Pose-Aware Signed Distance Field (PASDF), a novel framework that integrates 3D anomaly detection and repair by learning a continuous, pose-invariant shape representation. PASDF leverages a Pose Alignment Module for canonicalization and a SDF Network to dynamically incorporate pose, enabling implicit learning of high-fidelity anomaly repair templates from the continuous SDF. This facilitates precise pixel-level anomaly localization through an Anomaly-Aware Scoring Module. Crucially, the continuous 3D representation in PASDF extends beyond detection, facilitating in-situ anomaly repair. Experiments on Real3D-AD and Anomaly-ShapeNet demonstrate state-of-the-art performance, achieving high object-level AUROC scores of 80.2% and 90.0%, respectively. These results highlight the effectiveness of continuous geometric representations in advancing 3D anomaly detection and facilitating practical anomaly region repair. The code is available at https://github.com/ZZZBBBZZZ/PASDF to support further research.

Problem

• Detecting 3D point cloud anomalies under pose variations
• Overcoming geometric fidelity issues in patch-based methods
• Integrating anomaly detection and repair via continuous SDF

Innovation

• Pose-Aware Signed Distance Field (PASDF) framework
• Continuous pose-invariant 3D shape representation
• Anomaly-Aware Scoring Module for localization