LUMINA：带能量协调协议的多厂商乳腺X线摄影基准
LUMINA: A Multi-Vendor Mammography Benchmark with Energy Harmonization Protocol
针对多厂商、多能量采集造成的域偏移，提出仅在前景像素空间做对齐的"能量协调"方法，把图像映射到低能量参考域且保持病灶形态；这是一种模型无关的预处理/域对齐手段，可直接迁移到其他多中心医学影像基准的跨域评测与部署流程中。
原文依据 a foreground-only pixel-space alignment method ("energy harmonization") that maps images to a low-energy reference while preserving lesion morphology
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论公开可用的全视野数字乳腺X线摄影（FFDM）数据集在规模、临床标注和厂商多样性方面仍然有限，阻碍了鲁棒模型的开发。我们提出 LUMINA，一个经过整理的多厂商 FFDM 数据集，它显式编码采集能量与厂商元数据，以刻画现有基准中常被忽略的、具有临床意义的外观差异。该数据集包含来自 468 名患者的 1824 幅图像（960 例良性、864 例恶性），并附有经病理确诊的标签、BI-RADS 评估和乳腺密度标注。LUMINA 覆盖六套采集系统，同时包含高能量与低能量成像风格，从而能够系统分析由厂商和能量引起的域偏移。为应对这些差异，我们提出一种仅作用于前景的像素空间对齐方法（“能量协调”），将图像映射到低能量参考域，同时保持病灶形态。我们在三项临床相关任务上对 CNN 与 Transformer 模型进行了基准评测：诊断（良性 vs. 恶性）、BI-RADS 分类和密度估计。双视图模型一致优于单视图模型。EfficientNet-B0 在诊断任务上取得 93.54% 的 AUC，Swin-T 在密度预测上取得最佳宏平均 AUC 89.43%。能量协调在多种架构上均带来性能提升，并产生更聚焦的 Grad-CAM 响应。总体而言，LUMINA 提供了（1）一个厂商多样化的基准，以及（2）一个模型无关的能量协调框架，用于可靠且可部署的乳腺X线摄影 AI。
英文原文摘要 1519 字符
Publicly available full-field digital mammography (FFDM) datasets remain limited in size, clinical annotations, and vendor diversity, hindering the development of robust models. We introduce LUMINA, a curated, multi-vendor FFDM dataset that explicitly encodes acquisition energy and vendor metadata to capture clinically relevant appearance variations often overlooked in existing benchmarks. This dataset contains 1824 images from 468 patients (960 benign, 864 malignant), with pathology-confirmed labels, BI-RADS assessments, and breast-density annotations. LUMINA spans six acquisition systems and includes both high- and low-energy imaging styles, enabling systematic analysis of vendor- and energy-induced domain shifts. To address these variations, we propose a foreground-only pixel-space alignment method ("energy harmonization") that maps images to a low-energy reference while preserving lesion morphology. We benchmark CNN and transformer models on three clinically relevant tasks: diagnosis (benign vs. malignant), BI-RADS classification, and density estimation. Two-view models consistently outperform single-view models. EfficientNet-B0 achieves an AUC of 93.54% for diagnosis, while Swin-T achieves the best macro-AUC of 89.43% for density prediction. Harmonization improves performance across architectures and produces more localized Grad-CAM responses. Overall, LUMINA provides (1) a vendor-diverse benchmark and (2) a model-agnostic harmonization framework for reliable and deployable mammography AI.