让视觉语言模型区分正常与异常，实现零样本缺陷检测
Anomaly-Aware Vision-Language Adapters for Zero-Shot Anomaly Detection
- 双分支结构分别处理正常和异常模式，动态路由实现不对称特征变换
- 在MVTec-AD上达到93.5%图像AUROC，跨领域无需微调即适配医学影像
- 适合工业质检与医疗图像异常检测，尤其无标注异常样本场景
零样本异常检测旨在未见类别中识别缺陷而无需目标特定训练。现有方法对所有样本应用相同特征变换，忽略了正常数据分布紧凑、异常数据多样这一根本差异。本文提出AVA-DINO，一种基于冻结DINOv3视觉特征的异常感知视觉-语言适配框架，包含针对正常与异常模式的双专用分支。在辅助数据上联合训练时，通过文本引导的路由机制与显式路由正则化促进分支专业化。测试时仅需输入图像与预定义语言描述，动态融合双分支，实现非对称激活。该设计避免退化的均匀路由，支持上下文相关特征变换。在九个工业与医学基准上实验表明，性能达到领先水平，在MVTec-AD上取得93.5%图像AUROC，且在无领域微调下实现强跨域泛化。
原文摘要 · Abstract (English)
Zero-shot anomaly detection aims to identify defects in unseen categories without target-specific training. Existing methods usually apply the same feature transformation to all samples, treating normal and anomalous data uniformly despite their fundamentally asymmetric distributions, compact normals versus diverse anomalies. We instead exploit this natural asymmetry by proposing AVA-DINO, an anomaly-aware vision-language adaptation framework with dual specialized branches for normal and anomalous patterns that adapt frozen DINOv3 visual features. During training on auxiliary data, the two branches are learned jointly with a text-guided routing mechanism and explicit routing regularization that encourages branch specialization. At test time, only the input image and fixed, predefined language descriptions are used to dynamically combine the two branches, enabling an asymmetric activation. This design prevents degenerate uniform routing and allows context-specific feature transformations. Experiments across nine industrial and medical benchmarks demonstrate state-of-the-art performance, achieving 93.5% image-AUROC on MVTec-AD and strong cross-domain generalization to medical imaging without domain-specific fine-tuning. https://github.com/aqeeelmirza/AVA-DINO
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。