面向胸部X光片间期变化学习的时序反转
Temporal Inversion for Learning Interval Change in Chest X-Rays
把「图像对反转」当作监督信号来训练时序视觉-语言模型，这个思路可直接迁移到任何需要建模方向性变化的视觉任务（如变化检测、视频时序顺序判别、术前术后对比），并提供了一个可复用的顺序敏感性与一致性评测协议。
原文依据 uses temporal inversion, reversing image pairs, as a supervisory signal to enhance the sensitivity of existing temporal vision--language models to directional change
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论视觉-语言预训练的最新进展催生了强大的医学基础模型，但其中大多数模型孤立地分析影像，忽视了对比既往与当前图像以评估间期变化这一关键临床任务。对于胸部X光片（CXR），捕捉间期变化至关重要，因为放射科医生不仅要评估病灶的静态表现，还要评估其随时间演变的方式。我们提出TILA（时序反转感知学习与对齐），这是一个简单而有效的框架，它利用时序反转，即反转图像对的顺序，作为监督信号，以增强现有时序视觉-语言模型对方向性变化的敏感性。TILA在预训练、微调和推理三个阶段整合了反转感知目标，以对时序顺序的显式学习补充传统的外观建模。我们还提出了一个统一的评测协议，用于评估时序反转下的顺序敏感性与一致性，并引入MS-CXR-T_retrieval，这是一个通过通用协议构建的检索评测集，该协议可应用于任何时序CXR数据集。在公开数据集和真实医院队列上的实验表明，将TILA应用于多种现有架构时，它能持续提升进展分类性能和时序嵌入对齐效果。
英文原文摘要 1349 字符
Recent advances in vision-language pretraining have enabled strong medical foundation models, yet most analyze radiographs in isolation, overlooking the key clinical task of comparing prior and current images to assess interval change. For chest radiographs (CXRs), capturing interval change is essential, as radiologists must evaluate not only the static appearance of findings but also how they evolve over time. We introduce TILA (Temporal Inversion-aware Learning and Alignment), a simple yet effective framework that uses temporal inversion, reversing image pairs, as a supervisory signal to enhance the sensitivity of existing temporal vision--language models to directional change. TILA integrates inversion-aware objectives across pretraining, fine-tuning, and inference, complementing conventional appearance modeling with explicit learning of temporal order. We also propose a unified evaluation protocol to assess order sensitivity and consistency under temporal inversion, and introduce MS-CXR-T_retrieval, a retrieval evaluation set constructed through a general protocol that can be applied to any temporal CXR dataset. Experiments on public datasets and real-world hospital cohorts demonstrate that TILA consistently improves progression classification and temporal embedding alignment when applied to multiple existing architectures.