Sequential PatchCore：使用合成杂质进行表面检测的异常检测
Sequential PatchCore: Anomaly Detection for Surface Inspection using Synthetic Impurities
AI总结：
针对表面检测中的水渍等杂质与高分辨率训练内存瓶颈，提出合成水渍数据生成方法和 Sequential PatchCore 顺序核心集训练框架，验证合成预训练与真实数据微调的性能收益。
摘要：
表面杂质（例如水渍、指纹、贴纸）的出现是一个经常被提及的问题，会导致自动化视觉检测系统性能下降。与此同时，用于视觉表面检测的合成数据生成技术主要集中于生成完美样本和缺陷，而忽视了杂质。本研究强调了在生成合成数据时考虑杂质的重要性。我们提出了一种程序化方法，以在合成数据中加入具有照片级真实感的水渍。生成的合成数据集与真实数据集相对应，并进一步用于训练异常检测模型和研究水渍的影响。用于表面检测的高分辨率图像会在异常检测训练期间造成内存瓶颈。为解决这一问题，我们提出了 Sequential PatchCore——一种顺序构建核心集的方法，使使用消费级硬件训练大尺寸图像变得可行。这使我们能够利用在不同数据集版本上预训练的核心集进行迁移学习。我们的结果展示了使用合成数据预训练显式核心集异常模型的益处，以及使用真实数据微调核心集所带来的进一步性能提升。我们观察到杂质和标注歧义如何降低模型性能，并额外报告了按缺陷分类的召回率，以从工业相关视角呈现模型性能。
英文摘要：
The appearance of surface impurities (e.g., water stains, fingerprints, stickers) is an often-mentioned issue that causes degradation of automated visual inspection systems. At the same time, synthetic data generation techniques for visual surface inspection have focused primarily on generating perfect examples and defects, disregarding impurities. This study highlights the importance of considering impurities when generating synthetic data. We introduce a procedural method to include photorealistic water stains in synthetic data. The synthetic datasets are generated to correspond to real datasets and are further used to train an anomaly detection model and investigate the influence of water stains. The high-resolution images used for surface inspection lead to memory bottlenecks during anomaly detection training. To address this, we introduce Sequential PatchCore - a method to build coresets sequentially and make training on large images using consumer-grade hardware tractable. This allows us to perform transfer learning using coresets pre-trained on different dataset versions. Our results show the benefits of using synthetic data for pre-training an explicit coreset anomaly model and the extended performance benefits of finetuning the coreset using real data. We observed how the impurities and labelling ambiguity lower the model performance and have additionally reported the defect-wise recall to provide an industrially relevant perspective on model performance.
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。
