FiLo++: 通过融合细粒度描述和变形定位实现零/少样本异常检测
FiLo++: Zero-/Few-Shot Anomaly Detection by Fused Fine-Grained Descriptions and Deformable Localization
AI总结：
FiLo++通过融合细粒度描述和变形定位技术，提升零/少样本异常检测的准确性和鲁棒性。
摘要：
异常检测方法通常需要大量目标类的正常样本进行训练，限制了其在需要快速适应的场景中的应用，如冷启动。零样本和少样本异常检测不需要提前获取目标类的标记样本，因此成为有前景的研究方向。现有的零样本和少样本方法通常利用强大的多模态模型通过比较图像-文本相似性来检测和定位异常。然而，它们的手工通用描述无法捕捉不同对象中可能出现的多样化异常，简单的片段级图像-文本匹配往往难以定位形状和大小各异的异常区域。为了解决这些问题，本文提出了FiLo++方法，包括两个关键组件。第一个组件，融合细粒度描述（FusDes），利用大型语言模型为每个对象类别生成异常描述，结合固定和可学习的提示模板，并应用运行时提示过滤方法，生成更准确和任务特定的文本描述。第二个组件，变形定位（DefLoc），整合了视觉基础模型Grounding DINO与位置增强的文本描述和多尺度变形跨模态交互（MDCI）模块，能够准确定位各种形状和大小的异常。此外，我们设计了一种位置增强的片段匹配方法以提高少样本异常检测性能。在多个数据集上的实验表明，FiLo++相比现有方法取得了显著的性能提升。代码将在https://github.com/CASIA-IVA-Lab/FiLo上提供。
英文摘要：
Anomaly detection methods typically require extensive normal samples from the target class for training, limiting their applicability in scenarios that require rapid adaptation, such as cold start. Zero-shot and few-shot anomaly detection do not require labeled samples from the target class in advance, making them a promising research direction. Existing zero-shot and few-shot approaches often leverage powerful multimodal models to detect and localize anomalies by comparing image-text similarity. However, their handcrafted generic descriptions fail to capture the diverse range of anomalies that may emerge in different objects, and simple patch-level image-text matching often struggles to localize anomalous regions of varying shapes and sizes. To address these issues, this paper proposes the FiLo++ method, which consists of two key components. The first component, Fused Fine-Grained Descriptions (FusDes), utilizes large language models to generate anomaly descriptions for each object category, combines both fixed and learnable prompt templates and applies a runtime prompt filtering method, producing more accurate and task-specific textual descriptions. The second component, Deformable Localization (DefLoc), integrates the vision foundation model Grounding DINO with position-enhanced text descriptions and a Multi-scale Deformable Cross-modal Interaction (MDCI) module, enabling accurate localization of anomalies with various shapes and sizes. In addition, we design a position-enhanced patch matching approach to improve few-shot anomaly detection performance. Experiments on multiple datasets demonstrate that FiLo++ achieves significant performance improvements compared with existing methods. Code will be available at https://github.com/CASIA-IVA-Lab/FiLo.
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。
