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

TL;DR

This paper presents PASDF, a novel continuous 3D shape representation framework that improves anomaly detection and repair in point clouds by achieving pose invariance and high geometric fidelity, leading to state-of-the-art results.

Contribution

Introduction of PASDF, a pose-aware signed distance field framework that unifies 3D anomaly detection and repair with continuous, high-fidelity shape modeling.

Findings

Achieves object-level AUROC scores of 80.2% and 90.0% on Real3D-AD and Anomaly-ShapeNet.

Enables precise pixel-level anomaly localization and in-situ repair.

Demonstrates state-of-the-art performance in 3D anomaly detection tasks.

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.