面向无监督异常检测的多原型紧致性与边界感知合成
Multi-Prototype Compactness and Boundary-Aware Synthesis for Unsupervised Anomaly Detection
用多原型混合建模正常特征分布、并在簇间拓扑边界处合成伪异常来收紧决策面，这套「原型聚类+边界合成」思路可直接迁移到工业缺陷检测与开放集异常定位任务中，替代单超球假设。
原文依据 the Boundary-Aware Anomaly Synthesis (BAAS) module generates pseudo-anomalies at the topological boundaries between MPCC clusters
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论无监督异常检测（UAD）对工业质量控制至关重要。许多现有的基于嵌入的方法采用单原型假设，例如学习一个紧致超球以包围所有正常特征。然而，该策略在由光照、姿态或纹理变化引起的类内方差下会失效。为覆盖所有多样的正常样本，单一原型可能诱导出过于宽松的决策边界，使细微异常难以被检测。为克服这一局限，我们提出 PGBL（原型引导的边界学习），一个将结构化表征学习与针对性异常合成相协同的框架。首先，多原型紧致约束（MPCC）模块将正常特征分布建模为多个语义原型的混合，从而为每个正常子模式获得更紧致的局部表征。其次，边界感知异常合成（BAAS）模块并非盲目合成异常，而是在 MPCC 各簇之间的拓扑边界处生成伪异常。最后，判别式边界精炼器（DBR）通过将正常簇与合成异常分离，学习塑造最终的决策面。在 MVTec-AD、VisA 和 Real-IAD 上的大量实验表明，PGBL 在异常检测与定位两方面均持续优于先前方法。
英文原文摘要 1360 字符
Unsupervised Anomaly Detection (UAD) is crucial for industrial quality control. Many existing embedding-based methods adopt a single-prototype assumption and learn, for example, a compact hypersphere to enclose all normal features. However, this strategy breaks down under intra-class variance caused by changes in illumination, pose, or texture. To cover all diverse normal samples, a single prototype may induce an overly loose decision boundary, making subtle anomalies hard to detect. To overcome this limitation, we propose PGBL (Prototype-Guided Boundary Learning), a framework that synergizes structured representation learning with targeted anomaly synthesis. First, the Multi-Prototype Compact Constraint (MPCC) module models the normal feature distribution as a mixture of multiple semantic prototypes, enabling tighter local representations for each normal sub-pattern. Second, instead of blind anomaly synthesis, the Boundary-Aware Anomaly Synthesis (BAAS) module generates pseudo-anomalies at the topological boundaries between MPCC clusters. Finally, a Discriminative Boundary Refiner (DBR) learns to shape the final decision surface by separating normal clusters from the synthesized anomalies. Extensive experiments on MVTec-AD, VisA and Real-IAD show that PGBL consistently outperforms prior methods in both anomaly detection and localization.