arXiv 2505.233412026-06-09cs.CV

双流注意力引导学习用于弱监督全切片图像分类

Dual-stream attention-guided learning for weakly supervised whole slide image classification

College of Computer Science and Technology, College of Data Science, Taiyuan University of Technology（太原科技大学计算机科学与技术学院、数据科学学院）；College of Humanities, Law and Foreign Languages, Taiyuan University of Technology（太原科技大学人文学院、法律与外语学院）；College of Artificial Intelligence, Taiyuan University of Technology（太原科技大学人工智能学院）；School of Cyberspace Security, Beijing University of Posts and Telecommunications（北京邮电大学网络安全学院）；School of Mathematics, Taiyuan University of Technology（太原科技大学数学学院）

Daoxi Cao, Hangbei Cheng, Yijin Li, Ruolin Zhou, Xuehan Zhang, Xinyi Li, Binwei Li, Xuancheng Gu, Jianan Zhang, Xueyu Liu, Yongfei Wu

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

提出双流注意力引导学习框架，通过师生双流架构和注意力引导伪标签，解决弱监督下全切片图像中关键区域识别和实例关系建模问题，在合成和真实病理数据集上优于现有方法。

摘要：

全切片图像（WSIs）因其超高分辨率和丰富的形态学信息在癌症诊断中发挥关键作用，多实例学习（MIL）已成为解决WSIs巨大尺寸和实例细粒度标注稀缺的主流范式。然而，现有大多数MIL方法难以仅使用切片级标签准确识别诊断关键局部区域（实例），并且在高效建模实例间关系方面存在不足。为解决这些问题，我们提出了一种双流注意力引导学习（DSAGL）框架。DSAGL通过师生双流架构桥接切片级监督和实例级学习，并通过生成注意力引导伪标签缓解实例歧义。该框架采用共享轻量级编码器高效建模长距离依赖，并利用基于注意力的融合机制增强对稀疏信息区域的敏感性。在合成基准和真实病理WSI数据集上的大量实验表明，DSAGL在弱监督下始终优于最先进的MIL方法，实现了卓越的判别性能和鲁棒性。

英文摘要：

Whole slide images (WSIs) play a crucial role in cancer diagnosis due to their ultra-high resolution and rich morphological information, and multiple instance learning (MIL) has become a prevalent paradigm to solve the massive size of WSIs and the scarcity of fine-grained annotations of instance. However, most existing MIL methods struggle to accurately identify diagnostically critical local regions (instance) using only slide-level labels, and suffer from modelling the relationship of instances efficiently. To address these defects, we propose a Dual-Stream Attention-Guided Learning (DSAGL) framework. DSAGL bridges slide-level supervision and instance-level learning through a teacher-student dual-stream architecture, and mitigates instance ambiguity by generating attention-guided pseudo labels. The framework employs a shared lightweight encoder to efficiently model long-range dependencies and an attention-based fusion mechanism to enhance sensitivity to sparse, informative regions. Extensive experiments on synthetic benchmarks and real-world pathological WSI datasets demonstrate that DSAGL consistently outperforms state-of-the-art MIL methods, achieving superior discriminative performance and robustness under weak supervision.

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。