arXiv 2506.076522025-06-09cs.CVcs.AI

FMaMIL：用于医学图像弱监督病灶分割的频率驱动 Mamba 多示例学习

FMaMIL: Frequency-Driven Mamba Multi-Instance Learning for Weakly Supervised Lesion Segmentation in Medical Images

College of Computer Science and Technology (College of Data Science), Taiyuan University of Technology, Taiyuan, Shanxi, 030024, China（计算机科学与技术学院（数据科学学院），太原科技大学，太原，山西，030024，中国）；College of Artificial Intelligence, Taiyuan University of Technology, Taiyuan, Shanxi, 030024, China（人工智能学院，太原科技大学，太原，山西，030024，中国）；School of mathematics, Taiyuan University of Technology, Taiyuan, 030024, China（数学学院，太原科技大学，太原，030024，中国）；School of Artificial Intelligence, Beijing Normal University, Beijing, 100875, China（人工智能学院，北京师范大学，北京，100875，中国）；National Institute for Data Science in Health（健康数据科学国家研究院）；the Department of Computer Science, School of Informatics, Xiammen, 361005, China（信息学院计算机科学系，厦门，361005，中国）；School of Software, Dalian University of Technology, Dalian,116621, China（软件学院，大连理工大学，大连，116621，中国）

Hangbei Cheng, Xiaorong Dong, Xueyu Liu, Jianan Zhang, Xuetao Ma, Mingqiang Wei, Liansheng Wang, Junxin Chen, Yongfei Wu

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

本文提出两阶段弱监督框架 FMaMIL，利用 Mamba、频域编码、CAM 软标签监督与自校正机制，仅靠图像级标签实现组织病理病灶分割。

摘要：

组织病理学图像中的准确病灶分割对于诊断解释和定量分析至关重要，但由于昂贵的像素级标注数量有限，这一任务仍具挑战性。为解决该问题，我们提出 FMaMIL，这是一种仅基于图像级标签进行弱监督病灶分割的新型两阶段框架。第一阶段引入轻量级的基于 Mamba 的编码器，以在 MIL 范式下捕获图像块之间的长程依赖关系。为增强空间敏感性和结构感知能力，我们设计了可学习的频域编码模块，利用基于频谱的信息补充空间域特征。该阶段生成的 CAM 用于指导分割训练。第二阶段通过 CAM 引导的软标签监督和自校正机制细化初始伪标签，即使在标签噪声下也能实现鲁棒训练。在公开和私有组织病理学数据集上的大量实验表明，FMaMIL 在不依赖像素级标注的情况下优于最先进的弱监督方法，验证了其在数字病理学应用中的有效性和潜力。

英文摘要：

Accurate lesion segmentation in histopathology images is essential for diagnostic interpretation and quantitative analysis, yet it remains challenging due to the limited availability of costly pixel-level annotations. To address this, we propose FMaMIL, a novel two-stage framework for weakly supervised lesion segmentation based solely on image-level labels. In the first stage, a lightweight Mamba-based encoder is introduced to capture long-range dependencies across image patches under the MIL paradigm. To enhance spatial sensitivity and structural awareness, we design a learnable frequency-domain encoding module that supplements spatial-domain features with spectrum-based information. CAMs generated in this stage are used to guide segmentation training. In the second stage, we refine the initial pseudo labels via a CAM-guided soft-label supervision and a self-correction mechanism, enabling robust training even under label noise. Extensive experiments on both public and private histopathology datasets demonstrate that FMaMIL outperforms state-of-the-art weakly supervised methods without relying on pixel-level annotations, validating its effectiveness and potential for digital pathology applications.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。