发表机构
Concordia University; Université de Montréal; Mila - Quebec AI Institute; Centre de recherche du CHUM（康考迪亚大学; 蒙特利尔大学; 米拉-魁北克人工智能研究所; CHUM研究中心）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
针对病理Transformer自注意力的局部空间冗余失效模式，提出轻量即插即用的Gated SRP模块，在TCGA生存队列和切片分类数据集上均实现性能提升，仅增加极少量参数。
AI 中文摘要
Transformer模型正越来越多地被用于计算病理学中的全切片图像分析。然而，全切片图像（WSI）与自然图像存在根本差异：相邻的图像块常包含高度相似的组织类型、染色、纹理及细胞组成。我们将这种局部空间冗余识别为自注意力的一种病理特定失效模式，其中占主导的邻域特征会被重复混入图像块令牌中，削弱细微的诊断或预后偏差。我们提出了门控空间冗余投影（Gated SRP），这是一种用于自注意力层的轻量型即插即用修正模块。对于每个图像块令牌和注意力头，Gated SRP从相邻的值向量中估计局部冗余轴，将注意力输出投影到该轴上，并应用一个学习到的带符号门控，以从几何上修正与冗余对齐的分量。在5个TCGA生存队列中，Gated SRP在所有队列中均获得了所对比注意力变体中的最高平均C指数，相比基础注意力实现了平均提升，且仅增加了+0.02%的参数；在5个切片级分类数据集上，它在16个报告指标中的12个上优于基础注意力，并在3个数据集上取得了最佳AUC。代码可在该https URL公开获取。
英文摘要
Transformer models are increasingly used for whole-slide image analysis in computational pathology. Yet, WSIs differ fundamentally from natural images: neighbouring patches often contain highly similar tissue type, stain, texture, and cellular composition. We identify this local spatial redundancy as a pathology-specific failure mode of self-attention, where dominant neighbourhood features can be repeatedly mixed into patch-tokens and weaken subtle diagnostic or prognostic deviations. We propose Gated Spatial Redundancy Projection (Gated SRP), a lightweight drop-in correction module for self-attention layers. For each patch token and attention head, Gated SRP estimates a local redundancy axis from neighbouring value vectors, projects the attention output onto this axis, and applies a learned signed gate to correct the redundancy-aligned component geometrically. Across five TCGA survival cohorts, Gated SRP obtains the highest mean C-index among the compared attention variants in all cohorts, with an average improvement over the base attention, while adding only +0.02% parameters. Across five slide-level classification datasets, it improves the base attention on 12 of 16 reported metrics and achieves the best AUC on three datasets. Code is publicly available at https://github.com/AtlasAnalyticsLab/GatedSRP.
CommentsBMVC 2026 Oral Presentation
