基于数字孪生与上下文记忆的复杂度感知手术视频推理分割
原标题：Complexity-Aware Surgical Video Reasoning Segmentation with Digital Twin and Context Memory
AI 摘要
研究者提出 SurgReasonSeg，一种面向手术视频的复杂度感知推理分割方法：它基于手术基础模型构建数字孪生（DT）表示，并用强化学习训练同一大语言模型的两个 LoRA 适配器，分别处理不同复杂度的查询，再由同样用强化学习训练的轻量路由器在延迟与精度之间做选择。该方法还引入自适应滑动窗口和手术上下文记忆，按工作流相似度检索历史病例。作者同时发布 EndoVis-ReasonSeg 基准（1,460 条训练样本、961 条测试样本），在语义、空间、时序查询上分别比最强基线 ReSurgSAM2 提升 0.178、0.237、0.297，并以 18.8 FPS 实现实时推理分割。
正文节选
Complexity-Aware Surgical Video Reasoning Segmentation with Digital Twin and Context Memory Abstract Conventional surgical segmentation relies on closed-set categories or explicit text queries, which cannot resolve implicit queries that require reasoning about workflows, instrument functions, and temporal relationships. To address these limitations, we propose SurgReasonSeg, a complexity-aware method that brings reasoning segmentation over digital twin (DT) representations to surgical video. Spe