arXiv 2511.005402026-02-12cs.CV

现实工业异常检测数据集：推动工业异常检测数据集进入现代时代

Real-IAD Variety: Pushing Industrial Anomaly Detection Dataset to a Modern Era

Fudan University（复旦大学）；Youtu Lab, Tencent（腾讯优图实验室）；Shanghai Jiao Tong University（上海交通大学）；Rongcheer Co., Ltd（融雪科技有限公司）；City University of Hong Kong（香港城市大学）；National University of Singapore（新加坡国立大学）；Shanghai Ocean University（上海海洋大学）

Wenbing Zhu, Chengjie Wang, Bin-Bin Gao, Jiangning Zhang, Guannan Jiang, Jie Hu, Zhenye Gan, Lidong Wang, Ziqing Zhou, Jianghui Zhang, Linjie Cheng, Yurui Pan, … 展开作者

Wenbing Zhu, Chengjie Wang, Bin-Bin Gao, Jiangning Zhang, Guannan Jiang, Jie Hu, Zhenye Gan, Lidong Wang, Ziqing Zhou, Jianghui Zhang, Linjie Cheng, Yurui Pan, Bo Peng, Mingmin Chi, Lizhuang Ma

备注：: 17 pages, 8 figures and 7 tables

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

Real-IAD Variety 是一个大规模、高多样性的工业异常检测数据集，用于评估和训练下一代基础IAD模型，揭示了现有方法在类别扩展时的性能下降问题。

摘要：

工业异常检测（IAD）是确保操作安全、保持产品质量和优化制造效率的基石。然而，IAD算法的进步严重受到现有公开基准的限制。当前的数据集往往存在类别多样性有限和规模不足的问题，导致在复杂的真实世界场景中性能饱和和模型迁移性差。为弥合这一差距，我们引入了Real-IAD Variety，这是最大的且最多样化的IAD基准。它包含198,950张高分辨率图像，覆盖160个不同的物体类别。该数据集通过涵盖28个行业、24种材料类型、22种颜色变化和27种缺陷类型，确保了前所未有的多样性。我们的广泛实验分析突显了这一基准所带来的重大挑战：最先进的多类无监督异常检测方法在类别从30扩展到160时，性能显著下降（范围从10%到20%）。相反，我们证明零样本和少样本IAD模型在类别规模扩大时表现出显著的鲁棒性，保持了稳定的性能，并在多样化的工业环境中显著增强了泛化能力。这种前所未有的规模使Real-IAD Variety成为训练和评估下一代基础IAD模型的重要资源。

英文摘要：

Industrial Anomaly Detection (IAD) is a cornerstone for ensuring operational safety, maintaining product quality, and optimizing manufacturing efficiency. However, the advancement of IAD algorithms is severely hindered by the limitations of existing public benchmarks. Current datasets often suffer from restricted category diversity and insufficient scale, leading to performance saturation and poor model transferability in complex, real-world scenarios. To bridge this gap, we introduce Real-IAD Variety, the largest and most diverse IAD benchmark. It comprises 198,950 high-resolution images across 160 distinct object categories. The dataset ensures unprecedented diversity by covering 28 industries, 24 material types, 22 color variations, and 27 defect types. Our extensive experimental analysis highlights the substantial challenges posed by this benchmark: state-of-the-art multi-class unsupervised anomaly detection methods suffer significant performance degradation (ranging from 10% to 20%) when scaled from 30 to 160 categories. Conversely, we demonstrate that zero-shot and few-shot IAD models exhibit remarkable robustness to category scale-up, maintaining consistent performance and significantly enhancing generalization across diverse industrial contexts. This unprecedented scale positions Real-IAD Variety as an essential resource for training and evaluating next-generation foundation IAD models.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。