Open-Set Supervised 3D Anomaly Detection: An Industrial Dataset and a Generalisable Framework for Unknown Defects

Hanzhe Liang; Luocheng Zhang; Junyang Xia; HanLiang Zhou; Bingyang Guo; Yingxi Xie; Can Gao; Ruiyun Yu

cs.CVApr 1, 2026

Open-Set Supervised 3D Anomaly Detection: An Industrial Dataset and a Generalisable Framework for Unknown Defects

Authors: Hanzhe Liang, Luocheng Zhang, Junyang Xia, HanLiang Zhou, Bingyang Guo, Yingxi Xie, Can Gao, Ruiyun Yu,

Organizations: Mohamed bin Zayed University of Artificial Intelligence, Abu Dhabi, UAE · Shenzhen Audencia Financial Technology Institute, Shenzhen, China · School of Computer Science, Hangzhou Dianzi University, Hangzhou, China · Software College, Northeastern University, Shenyang, China ·

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.

Figures & tables

Fig. 1: Comparison of 3D anomaly detection settings. Prior methods are unsupervised or rely on synthetic negatives without real negative-sample constraints, the model lacks a clear optimization direction and has no reliable notion of the negative feature distribution, leading to weak anomaly awareness. propose open-set supervised 3D anomaly detection, where the training set includes a few real anomalies from a subset of categories, enabling the model to learn a grounded notion of the negative feature distribution, while testing requires detecting unseen anomaly categories.

Dataset Comparison
Dataset	Class	Synthetic	Anomaly Types	Ratio
MvTec3D-AD [ 2 ]	10	Real	3	$\sim$ 2%
Real3D-AD [ 33 ]	12	Real	2	$\sim$ 3%
Anomaly-ShapeNet [ 23 ]	40	Syn	4	$\sim$ 2%
MulSen-AD [ 24 ]	15	Real	5	$\sim$ 4%
IEC3D-AD [ 16 ]	15	Real	2	$\sim$ 1%

TABLE I: Comparison of representative 3D anomaly detection datasets and learning settings. For fairness, only shared categories across datasets are compared. For the multimodal benchmarks MVTec 3D-AD and MulSen-AD, only anomalies observable in point clouds are considered. As IEC3D-AD is not open source, we contacted the author to obtain details.

Fig. 2: Overview of the Open-Industry, comprising 15 categories, each containing five different anomalies. The red box represents scar, the green box represents scratch, the purple box represents concave, the brown box represents deformation, and the yellow box represents convex. The acquisition system and communication interaction are shown on the right.

Fig. 3: Workflow of our Open3D-AD. It consists of three steps: (i) anomaly synthesis and point-wise feature encoding to form normal/anomalous mixture representations, (ii) greedily subsample the normal support and use it to refine the anomalous support via deconfounding, and (iii) dual-distribution aggregated scoring for anomaly detection.

Method $\rightarrow$	PatchCore			BTF
Metrics $\downarrow$	FPFH	FPFH+Raw	PointMAE	Raw	FPFH
O-AUROC	51.78	55.63	50.77	93.15	49.40
P-AUROC	54.34	53.14	50.72	66.31	51.14
O-AUPRC	52.45	56.90	53.01	94.15	51.41
P-AUPRC	00.70	00.97	01.30	01.11	00.72

TABLE VI: Performance comparison of self-supervised methods. All results are reported in the Supplementary Materials .

Method	Time Complexity	O-AUROC	P-AUROC	FPS
$M_{1}$	$\mathcal{O}(mN)$	84.39 $\pm$ 0.44	70.52 $\pm$ 0.56	8.4
$M_{2}$	$\mathcal{O}(1)$	80.10 $\pm$ 1.32	67.43 $\pm$ 2.40	10.1
$M_{3}$	$\mathcal{O}(N)$	81.33 $\pm$ 1.88	68.98 $\pm$ 2.18	9.6
$M_{4}$	$\mathcal{O}(N^{2})$	83.20 $\pm$ 0.95	69.84 $\pm$ 2.21	8.8
$M_{5}$	$\mathcal{O}(mN)$	83.50 $\pm$ 4.69	69.66 $\pm$ 6.35	9.0

TABLE VII: Performance comparison of different subsampling strategies on Open-Industry under 5 samples supervised.

Method	O-AUROC	P-AUROC
$M_{6}$	83.08 $\pm$ 0.59	69.18 $\pm$ 0.40
$M_{7}$	83.16 $\pm$ 0.29	70.17 $\pm$ 0.79
$M_{8}$	83.21 $\pm$ 0.58	70.25 $\pm$ 0.24
$M_{9}$	84.39 $\pm$ 0.44	70.52 $\pm$ 0.56

TABLE VIII: Ablation study on Open-Industry under the 5 samples per-class supervision setting.

Fig. 4: Feature distribution visualizations in the training and inference stages. A and B show selected training features of Normal, Seen Abnormal, and Simulated Abnormal. C and D show selected test features of Normal, Seen Abnormal, and Unseen Abnormal. SS denotes the Silhouette Score. In A and B, $SS_{1}$ , $SS_{2}$ , and $SS_{3}$ correspond to Normal vs. Seen Abnormal, Normal vs. Simulated Abnormal, and Seen Abnormal vs. Simulated Abnormal, respectively. In C and D, they correspond to Normal vs. Seen Abnormal, Normal vs. Unseen Abnormal, and Seen Abnormal vs. Unseen Abnormal, respectively.

Point Cloud Anomaly Detection

Point cloud anomaly detection identifies and localizes unusual 3D geometry or surface defects, often in industrial inspection with few labeled failures.

See all 25 papers

Synthesis4AD: Synthetic Anomalies are All You Need for 3D Anomaly Detection
Apr 6, 2026Yihan Sun, Yuqi Cheng, Junjie Zu +5Point Cloud Anomaly Detection Industrial Anomaly Detection
the National Center of Technology Innovation for Intelligent Design and Numerical Control, Huazhong University of Science and Technology, Wuhan 430074, China · the Department of Intelligent Manufacturing, Contemporary Amperex Technology Ltd., Ningde 352000, China · the Institute for Infocomm Research, A*STAR, Singapore 138632 +2
Physics-inspired Pseudo Anomaly Generation and Prototype Feature Guidance for 3D Anomaly Detection
Jul 12, 2026Jian Ning, Qin Zou, Linchun Wu +4Point Cloud Anomaly Detection Industrial Anomaly Detection
School of Computer Science, Wuhan University, Wuhan 430072, Hubei, China · Guangdong Laboratory of Artificial Intelligence and Digital Economy (SZ), Shenzhen 518060, Guangdong, China · School of Information and Software Engineering, East China Jiaotong University, Nanchang 330013, Jiangxi, China +1
AT3D-AD: Anomaly Type-Aware 3D Anomaly Detection via Hierarchical Point-Language Alignment
Sep 22, 2026Jingyu Zeng, Haoquan Lu, Can GaoPoint Cloud Anomaly Detection Industrial Anomaly Detection
College of Computer Science and Software Engineering and the Guangdong Key Laboratory of Intelligent Information Processing, Shenzhen University, Shenzhen 518060, China

Apr 6, 2026J/K move · Enter open · S save

Synthesis4AD: Synthetic Anomalies are All You Need for 3D Anomaly Detection

Yihan Sun, Yuqi Cheng, Junjie Zu, Yuxiang Tan+4

the National Center of Technology Innovation for Intelligent Design and Numerical Control, Huazhong University of Science and Technology, Wuhan 430074, China · the Department of Intelligent Manufacturing, Contemporary Amperex Technology Ltd., Ningde 352000, China · the Institute for Infocomm Research, A*STAR, Singapore 138632+2

Point Cloud Anomaly Detection Industrial Anomaly Detection Synthetic Anomaly Generation

Open paper AskPDF arXiv

Industrial 3D anomaly detection performance is fundamentally constrained by the scarcity and long-tailed distribution of abnormal samples. To address this challenge, we propose Synthesis4AD, an end-to-end paradigm that leverages large-scale, high-fidelity synthetic anomalies to learn more discriminative representations for 3D anomaly detection. At the core of Synthesis4AD is 3D-DefectStudio, a software platform built upon the controllable synthesis engine MPAS, which injects geometrically realistic defects guided by higher-dimensional support primitives while simultaneously generating accurate point-wise anomaly masks. Furthermore, Synthesis4AD incorporates a multimodal large language model (MLLM) to interpret product design information and automatically translate it into executable anomaly synthesis instructions, enabling scalable and knowledge-driven anomalous data generation. To improve the robustness and generalization of the downstream detector on unstructured point clouds, Synthesis4AD further introduces a training pipeline based on spatial-distribution normalization and geometry-faithful data augmentations, which alleviates the sensitivity of Point Transformer architectures to absolute coordinates and improves feature learning under realistic data variations. Extensive experiments demonstrate state-of-the-art performance on Real3D-AD, MulSen-AD, and a real-world industrial parts dataset. The proposed synthesis method MPAS and the interactive system 3D-DefectStudio will be publicly released at https://github.com/hustCYQ/Synthesis4AD.

Abstract

Point Cloud Anomaly Detection

Explore similar work

Synthesis4AD: Synthetic Anomalies are All You Need for 3D Anomaly Detection

Physics-inspired Pseudo Anomaly Generation and Prototype Feature Guidance for 3D Anomaly Detection

AT3D-AD: Anomaly Type-Aware 3D Anomaly Detection via Hierarchical Point-Language Alignment