Boosting Global-Local Feature Matching via Anomaly Synthesis for Multi-Class Point Cloud Anomaly Detection

2025-05-12

IEEE Transactions on Automation Science and Engineering

AI Summary

In multi-class point cloud unsupervised anomaly detection, performance degrades due to feature confusion between normal and anomalous samples across categories. To address this, we propose a global-local feature matching framework. Our contributions are threefold: (1) a novel anomaly-synthesis-based data augmentation strategy that enhances the feature extractor’s discriminative capability for fine-grained anomalies; (2) a decoupled dual-granularity memory bank—separating intra-class and cross-class feature distributions explicitly to mitigate confusion; and (3) a unified multi-class detection mechanism requiring no anomalous samples, enabling end-to-end unsupervised scoring via adaptive distance metrics. Evaluated on MVTec 3D-AD, Real3D-AD, and real-world industrial datasets, our method achieves state-of-the-art performance, significantly improving detection accuracy and generalization in complex multi-class mixed scenarios.

Technology Category

• Machine Learning: Multi-class/Multi-label Learning & Extreme Classification
• Data Mining & Knowledge Management: Anomaly/Outlier Detection
• Computer Vision: Object Detection & Categorization

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for heterogeneous, signed, attributed, multi-relational, temporal, higher-order, and annotated Web-related graphs
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets

Abstract

Point cloud anomaly detection is essential for various industrial applications. The huge computation and storage costs caused by the increasing product classes limit the application of single-class unsupervised methods, necessitating the development of multi-class unsupervised methods. However, the feature similarity between normal and anomalous points from different class data leads to the feature confusion problem, which greatly hinders the performance of multi-class methods. Therefore, we introduce a multi-class point cloud anomaly detection method, named GLFM, leveraging global-local feature matching to progressively separate data that are prone to confusion across multiple classes. Specifically, GLFM is structured into three stages: Stage-I proposes an anomaly synthesis pipeline that stretches point clouds to create abundant anomaly data that are utilized to adapt the point cloud feature extractor for better feature representation. Stage-II establishes the global and local memory banks according to the global and local feature distributions of all the training data, weakening the impact of feature confusion on the establishment of the memory bank. Stage-III implements anomaly detection of test data leveraging its feature distance from global and local memory banks. Extensive experiments on the MVTec 3D-AD, Real3D-AD and actual industry parts dataset showcase our proposed GLFM’s superior point cloud anomaly detection performance.Note to Practitioners—The proposed GLFM is employed for point cloud anomaly detection in industrial inspection, capable of simultaneously processing data across multiple classes. GLFM requires the collection of a set of normal product samples for model training, where the features of these samples are stored. If the feature distribution of a test sample deviates substantially from that of the normal samples, it is flagged as anomalous. GLFM not only exhibits outstanding performance on public datasets but has also been validated on a real-world industrial parts point cloud dataset.

Problem

• Develop multi-class unsupervised point cloud anomaly detection
• Address feature confusion between normal and anomalous points
• Enhance detection via global-local feature matching stages

Innovation

• Anomaly synthesis pipeline for better feature representation
• Global-local memory banks to reduce feature confusion
• Feature distance-based anomaly detection for test data