arXiv 2603.215622026-03-23cs.CV

探索多模态提示用于无监督连续异常检测

Exploring Multimodal Prompts For Unsupervised Continuous Anomaly Detection

Key Laboratory of Computing Power Network（计算能力网络重点实验室）；Information Security, Ministry of Education, Shandong Computer Science Center (National Supercomputer Center in Jinan), Qilu University of Technology (Shandong Academy of Sciences)（信息安全，教育部，山东计算机科学中心（济南国家超算中心），齐鲁工业大学（山东科学院））；Shandong Provincial Key Laboratory of Computing Power Internet（山东省计算能力互联网重点实验室）；Service Computing, Shandong Fundamental Research Center for Computer Science（服务计算，山东省计算机基础研究中心）；Faculty of Data Science, City University of Macau（数据科学学院，澳门城市大学）

Mingle Zhou, Jiahui Liu, Jin Wan, Gang Li, Min Li

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

本文提出基于多模态提示的无监督连续异常检测框架，通过持续多模态提示记忆库和缺陷语义引导自适应融合机制提升检测精度与鲁棒性，实验表明在MVTec AD和VisA数据集上取得最优性能。

摘要：

无监督连续异常检测（UCAD）因其能有效解决传统无监督异常检测（UAD）中灾难性遗忘和计算负担重的问题而受到关注。然而，现有仅依赖视觉信息的UCAD方法无法捕捉复杂场景中的正常 manifold，从而限制了异常检测精度的进一步提升。为克服这一限制，我们提出一个基于多模态提示的无监督连续异常检测框架。具体来说，我们引入了一个持续多模态提示记忆库（CMPMB），该记忆库在连续任务中逐步蒸馏并保留来自视觉和文本领域的原型正常模式，从而获得更丰富的正常性表示。此外，我们设计了一种缺陷语义引导自适应融合机制（DSG-AFM），该机制结合了自适应归一化模块（ANM）和动态融合策略（DFS），以共同提升检测精度和对抗鲁棒性。在MVTec AD和VisA数据集上的基准实验表明，我们的方法在图像级AUROC和像素级AUPR指标上均取得最佳性能。

英文摘要：

Unsupervised Continuous Anomaly Detection (UCAD) is gaining attention for effectively addressing the catastrophic forgetting and heavy computational burden issues in traditional Unsupervised Anomaly Detection (UAD). However, existing UCAD approaches that rely solely on visual information are insufficient to capture the manifold of normality in complex scenes, thereby impeding further gains in anomaly detection accuracy. To overcome this limitation, we propose an unsupervised continual anomaly detection framework grounded in multimodal prompting. Specifically, we introduce a Continual Multimodal Prompt Memory Bank (CMPMB) that progressively distills and retains prototypical normal patterns from both visual and textual domains across consecutive tasks, yielding a richer representation of normality. Furthermore, we devise a Defect-Semantic-Guided Adaptive Fusion Mechanism (DSG-AFM) that integrates an Adaptive Normalization Module (ANM) with a Dynamic Fusion Strategy (DFS) to jointly enhance detection accuracy and adversarial robustness. Benchmark experiments on MVTec AD and VisA datasets show that our approach achieves state-of-the-art (SOTA) performance on image-level AUROC and pixel-level AUPR metrics.

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。