AI summaryⓘ
The authors study a type of 3D anomaly detection used in manufacturing, where only normal data and a few known defective samples are available for training, but the goal is to find new, unknown defects. They created Open-Industry, a detailed dataset with different kinds of real anomalies from industrial production. The authors improved existing methods for 3D point cloud data and proposed Open3D-AD, which models both normal and anomalous data distributions using normal samples, simulated anomalies, and partial real anomalies. They also introduced a technique called Correspondence Distributions Subsampling to better separate normal from abnormal data. Their experiments showed that Open3D-AD works well on Open-Industry and other 3D anomaly datasets, highlighting the promise of this open-set supervised approach.
3D anomaly detectionopen-set learningpoint cloudsupervised learninganomaly detection datasetprobability density estimationsimulated anomaliesreal anomaliesdistribution modelingmanufacturing defect detection
Authors
Hanzhe Liang, Luocheng Zhang, Junyang Xia, HanLiang Zhou, Bingyang Guo, Yingxi Xie, Can Gao, Ruiyun Yu, Jinbao Wang, Pan Li
Abstract
Although self-supervised 3D anomaly detection assumes that acquiring high-precision point clouds is computationally expensive, in real manufacturing scenarios it is often feasible to collect a limited number of anomalous samples. Therefore, we study open-set supervised 3D anomaly detection, where the model is trained with only normal samples and a small number of known anomalous samples, aiming to identify unknown anomalies at test time. We present Open-Industry, a high-quality industrial dataset containing 15 categories, each with five real anomaly types collected from production lines. We first adapt general open-set anomaly detection methods to accommodate 3D point cloud inputs better. Building upon this, we propose Open3D-AD, a point-cloud-oriented approach that leverages normal samples, simulated anomalies, and partially observed real anomalies to model the probability density distributions of normal and anomalous data. Then, we introduce a simple Correspondence Distributions Subsampling to reduce the overlap between normal and non-normal distributions, enabling stronger dual distributions modeling. Based on these contributions, we establish a comprehensive benchmark and evaluate the proposed method extensively on Open-Industry as well as established datasets including Real3D-AD and Anomaly-ShapeNet. Benchmark results and ablation studies demonstrate the effectiveness of Open3D-AD and further reveal the potential of open-set supervised 3D anomaly detection.