3D inspection system improves fault detection with precise location and type
AT3D-AD: Anomaly Type-Aware 3D Anomaly Detection via Hierarchical Point-Language Alignment
Computer Vision and Pattern Recognition
Summary
Finding defects in 3D scans of objects is hard because it's often unclear exactly where the problem is and what kind it is. The authors created a new method called AT3D-AD that makes fake defects to learn from, then looks at the overall shape and small details together. It can spot, find, and name different kinds of problems better than earlier tools. Their tests show it works very well on several 3D datasets used for checking objects.
What this means in practice
- •For quality control engineers: Detect and locate different types of 3D surface defects on manufactured parts using improved anomaly recognition and classification.
- •For automated visual inspection developers: Develop inspection systems that generate training anomalies and align multi-scale features to improve defect detection precision on industrial 3D scans.
Authors
Jingyu Zeng, Haoquan Lu, Can Gao
Abstract
Detecting and localizing 3D point-cloud defects is essential for industrial inspection. However, existing methods often suffer from imprecise localization due to the lack of anomaly supervision and reliance on single-granularity representations. To address these limitations, we propose Anomaly Type-Aware 3D Anomaly Detection (AT3D-AD), a unified framework for joint detection, localization, and classification. Specifically, we first design the Physics-Driven Parametric Anomaly Synthesis (PDPAS) module employing multiple parametric functions to generate synthetic anomalies, providing explicit anomaly supervision. Then, we propose the Hierarchical Global-Local Anomaly Alignment (HiGLA) module to align global and local representations within the normal and anomalous groups. Finally, we propose the Semantic-Geometric Anomaly Classification (SGAC) module to jointly learn localization and classification, yielding spatially precise and type-discriminative anomaly representations. Extensive experiments establish new state-of-the-art performance on all four benchmarks. AT3D-AD achieves Object/Point AUROC scores of 98.1\%/98.9\% on Anomaly-ShapeNet and 95.0\%/95.2\% on Real3D-AD, while reaching 74.2\% Macro-F1 for anomaly-type recognition on Real3D-AD.