[论文] Re-thinking Mammography Transfer Learning: The Dataset-Informed Transf...

论文概要

研究领域: ML
作者: Adarsh Bhandary Panambur, Siming Bayer, Andreas Maier
发布时间: 2026-07-28
arXiv: 2607.26043

中文摘要

提升乳腺X线影像分类性能在小型精选数据集和大型临床队列中都是一个持续挑战。传统迁移学习方法通常忽略数据集特定特征，而近期的邻域知情方法局限于任务范围狭窄且形式僵化，难以扩展到人群水平数据集。为此，我们提出了数据集知情迁移学习（DITL）框架，将数据集衍生的难度信号与基于邻域的三元组监督整合到统一目标中。DITL引入了两个自适应组件：（i）自适应难度加权交叉熵（A-DWCE），根据自监督特征空间中k近邻标签纯度分配逐样本权重；（ii）自适应邻域表征三元组（A-NR-Triplet），使用可学习边距强制类内紧凑和类间分离。与焦点损失不同，DITL无需超参数调优，消除了启发式加权和固定边距，计算开销可忽略，是一种稳健且可扩展的优化策略。在大型VinDR-Mammo数据集上，DITL在全图像乳腺密度分类上达到最先进的性能，在准确率、F1分数和AUC上均有显著提升（p < 0.0001）。除了大型队列，DITL在小型ROI数据集上也提供了一致且具有统计显著性的提升（p < 0.0001）。通过连接小规模病变分析与大规模密度估计，DITL建立了一个临床相关、可扩展且可推广的乳腺X线分类框架，覆盖乳腺癌筛查到诊断的全谱系。

原文摘要

Enhancing classification performance in mammography remains a persistent challenge across both small curated datasets and large-scale clinical cohorts. Conventional transfer learning approaches often neglect dataset-specific characteristics, while recent neighborhood-informed methods have been restricted to narrow tasks with rigid formulations, limiting their scalability to population-level datasets. To address these challenges, we propose the Dataset-Informed Transfer Learning (DITL) framework, which integrates dataset-derived difficulty signals with neighborhood-based triplet supervision in a unified objective. DITL introduces two adaptive components: (i) Adaptive Difficulty-Weighted Cross-Entropy (A-DWCE), which assigns per-sample weights based on k-nearest neighbor label purity in a se...

自动采集于 2026-07-30

#论文 #arXiv #ML #小凯