AG-VAS：基于大语言多模态模型的锚点引导零样本视觉异常分割
AG-VAS: Anchor-Guided Zero-Shot Visual Anomaly Segmentation with Large Multimodal Models
通过可学习的语义锚点token（[SEG]、[NOR]、[ANO]）将抽象异常概念转化为空间定位实体，并设计语义-像素对齐模块（SPAM）与锚点引导掩码解码器（AGMD），为LMM-based零样本分割提供可复用的锚点引导范式。
原文依据 we present AG-VAS (Anchor-Guided Visual Anomaly Segmentation), a new framework that expands the LMM vocabulary with three learnable semantic anchor tokens
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论大语言多模态模型（LMMs）展现出强大的任务泛化能力，为零样本视觉异常分割（ZSAS）提供了新机遇。然而，现有的基于LMM的分割方法仍面临根本性局限：异常概念本质上是抽象且依赖上下文的，缺乏稳定的视觉原型；高层语义嵌入与像素级空间特征之间的弱对齐阻碍了精确的异常定位。为解决这些挑战，我们提出AG-VAS（锚点引导视觉异常分割），一种新框架，通过三个可学习的语义锚点token——[SEG]、[NOR]和[ANO]——扩展LMM词汇表，建立统一的锚点引导分割范式。具体而言，[SEG]作为绝对语义锚点，将抽象的异常语义转化为明确、空间接地的视觉实体（如孔洞或划痕），而[NOR]和[ANO]作为相对锚点，跨类别建模正常与异常模式之间的上下文对比。为进一步增强跨模态对齐，我们引入语义-像素对齐模块（SPAM），将语言级语义嵌入与高分辨率视觉特征对齐，并设计锚点引导掩码解码器（AGMD），执行锚点条件化的掩码预测以实现精确的异常定位。此外，我们构建了Anomaly-Instruct20K，一个大规模指令数据集，将异常知识组织为外观、形状和空间属性的结构化描述，促进所提出的语义锚点的有效学习与集成。在六个工业和医学基准上的大量实验表明，AG-VAS在零样本设置下取得了持续的最优性能。
英文原文摘要 1803 字符
Large multimodal models (LMMs) exhibit strong task generalization capabilities, offering new opportunities for zero-shot visual anomaly segmentation (ZSAS). However, existing LMM-based segmentation approaches still face fundamental limitations: anomaly concepts are inherently abstract and context-dependent, lacking stable visual prototypes, and the weak alignment between high-level semantic embeddings and pixel-level spatial features hinders precise anomaly localization. To address these challenges, we present AG-VAS (Anchor-Guided Visual Anomaly Segmentation), a new framework that expands the LMM vocabulary with three learnable semantic anchor tokens--[SEG], [NOR], and [ANO], establishing a unified anchor-guided segmentation paradigm. Specifically, [SEG] serves as an absolute semantic anchor that translates abstract anomaly semantics into explicit, spatially grounded visual entities (e.g., holes or scratches), while [NOR] and [ANO] act as relative anchors that model the contextual contrast between normal and abnormal patterns across categories. To further enhance cross-modal alignment, we introduce a Semantic-Pixel Alignment Module (SPAM) that aligns language-level semantic embeddings with high-resolution visual features, along with an Anchor-Guided Mask Decoder (AGMD) that performs anchor-conditioned mask prediction for precise anomaly localization. In addition, we curate Anomaly-Instruct20K, a large-scale instruction dataset that organizes anomaly knowledge into structured descriptions of appearance, shape, and spatial attributes, facilitating effective learning and integration of the proposed semantic anchors. Extensive experiments on six industrial and medical benchmarks demonstrate that AG-VAS achieves consistent state-of-the-art performance in the zero-shot setting.