SegMoTE：面向医学图像分割的 token 级混合专家
SegMoTE: Token-Level Mixture of Experts for Medical Image Segmentation
把 SAM 类通用分割基础模型迁移到医学影像时，可用 token 级混合专家加少量可学习参数实现模态/解剖自适应，并用渐进式提示词 token 化实现全自动分割，为低标注成本下的基础模型领域适配提供可复用范式。
原文依据 introducing only a small number of learnable parameters to dynamically adapt across modalities and tasks
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论医学图像分割对临床诊断与定量分析至关重要，但由于成像模态的异质性以及像素级标注成本高昂，该任务仍具挑战性。尽管 SAM 等通用交互式分割模型已取得显著进展，但其向医学影像的迁移仍面临两个关键瓶颈：(i) 缺乏面向模态与解剖结构特定任务的自适应机制，限制了其在分布外医学场景中的泛化能力；(ii) 现有的医学适配方法在大型异构数据集上不加选择地微调，导致监督信号含噪、成本更高以及负迁移。为解决这些问题，我们提出 SegMoTE，一个高效且自适应的医学图像分割框架。SegMoTE 保留 SAM 原有的提示词接口、高效推理与零样本泛化能力，同时仅引入少量可学习参数以在模态与任务间动态自适应。此外，我们设计了渐进式提示词 token 化机制，使全自动分割成为可能，显著降低对标注的依赖。在 MedSeg-HQ（一个规模不足现有大规模数据集 1% 的精选数据集）上训练后，SegMoTE 在多种成像模态与解剖任务上取得 SOTA 性能。它代表了在极低标注成本下将通用分割模型高效、鲁棒且可扩展地适配到医学领域的首次尝试，推动了基础视觉模型在临床应用中的实际部署。
英文原文摘要 1587 字符
Medical image segmentation is vital for clinical diagnosis and quantitative analysis, yet remains challenging due to the heterogeneity of imaging modalities and the high cost of pixel-level annotations. Although general interactive segmentation models like SAM have achieved remarkable progress, their transfer to medical imaging still faces two key bottlenecks: (i) the lack of adaptive mechanisms for modality- and anatomy-specific tasks, which limits generalization in out-of-distribution medical scenarios; and (ii) current medical adaptation methods fine-tune on large, heterogeneous datasets without selection, leading to noisy supervision, higher cost, and negative transfer. To address these issues, we propose SegMoTE, an efficient and adaptive framework for medical image segmentation. SegMoTE preserves SAM's original prompt interface, efficient inference, and zero-shot generalization while introducing only a small number of learnable parameters to dynamically adapt across modalities and tasks. In addition, we design a progressive prompt tokenization mechanism that enables fully automatic segmentation, significantly reducing annotation dependence. Trained on MedSeg-HQ, a curated dataset less than 1% of existing large-scale datasets, SegMoTE achieves SOTA performance across diverse imaging modalities and anatomical tasks. It represents the first efficient, robust, and scalable adaptation of general segmentation models to the medical domain under extremely low annotation cost, advancing the practical deployment of foundation vision models in clinical applications.