PDD：面向医学异常检测的流形先验多样性蒸馏
PDD: Manifold-Prior Diverse Distillation for Medical Anomaly Detection
把双教师（VMamba 全局上下文 + wide-ResNet50 局部结构）先验统一到共享高维流形再蒸馏给双学生，并用多样性损失防表征坍缩；这套「流形匹配统一 + 双学生互补蒸馏」可直接迁移到工业/医学以外的无监督异常检测与多源特征融合场景。
原文依据 unifies dual-teacher priors into a shared high-dimensional manifold and distills this knowledge into dual students with complementary behaviors
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论由于细微、异质的异常嵌入在复杂的解剖结构中，医学图像异常检测面临独特挑战。通过系统的 Grad-CAM 分析，我们发现判别性激活图在医学数据上失效，这与它们在工业数据集上的成功形成对比，从而凸显了流形级建模的必要性。我们提出 PDD（流形先验多样性蒸馏），这是一个新颖的框架，它将双教师先验统一到共享的高维流形中，并将该知识蒸馏到具有互补行为的双学生中。具体而言，冻结的 VMamba-Tiny 与 wide-ResNet50 编码器分别提供全局上下文先验与局部结构先验。它们的特征通过流形匹配与统一（MMU）模块进行统一，而层间特征适配（InA）模块则丰富中间表示。统一后的流形被蒸馏到两个学生中：一个通过 InA 执行逐层蒸馏以获得局部一致性，另一个通过流形先验仿射（MPA）模块接收跳跃投影表示以捕获跨层依赖关系。多样性损失可防止表征坍缩，同时保持检测灵敏度。在多个医学数据集上的大量实验表明，PDD 显著优于现有最优方法，在 HeadCT、BrainMRI 和 ZhangLab 数据集上，AUROC 分别提升高达 11.8%、2.9% 和 8.5%，在 Uni-Medical 数据集上 F1 max 提升 3.4%，确立了医学图像异常检测的新最优性能。实现代码将发布于 https://github.com/OxygenLu/PDD。
英文原文摘要 1686 字符
Medical image anomaly detection faces unique challenges due to subtle, heterogeneous anomalies embedded in complex anatomical structures. Through systematic Grad-CAM analysis, we reveal that discriminative activation maps fail on medical data, unlike their success on industrial datasets, motivating the need for manifold-level modeling. We propose PDD (Manifold-Prior Diverse Distillation), a novel framework that unifies dual-teacher priors into a shared high-dimensional manifold and distills this knowledge into dual students with complementary behaviors. Specifically, frozen VMamba-Tiny and wide-ResNet50 encoders provide global contextual and local structural priors, respectively. Their features are unified through a Manifold Matching and Unification (MMU) module, while an Inter-Level Feature Adaption (InA) module enriches intermediate representations. The unified manifold is distilled into two students: one performs layer-wise distillation via InA for local consistency, while the other receives skip-projected representations through a Manifold Prior Affine (MPA) module to capture cross-layer dependencies. A diversity loss prevents representation collapse while maintaining detection sensitivity. Extensive experiments on multiple medical datasets demonstrate that PDD significantly outperforms existing state-of-the-art methods, achieving improvements of up to 11.8%, 2.9%, and 8.5% in terms of AUROC on HeadCT, BrainMRI, and ZhangLab datasets, respectively, and 3.4% in terms of F1 max on the Uni-Medical dataset, establishing new state-of-the-art performance in medical image anomaly detection. The implementation will be released at https://github.com/OxygenLu/PDD.