arXiv 2603.199252026-06-23eess.IVcs.CV

ReconMIL: 协同潜在空间重建与双流Mamba的全切片图像分析

ReconMIL: Synergizing Latent Space Reconstruction with Bi-Stream Mamba for Whole Slide Image Analysis

USTC（中国科学技术大学）；NUDT（南京理工大学）；SCNU（华南师范大学）；NTU（南洋理工大学）；Anhui Province Key Laboratory of Biomedical Imaging and Intelligent Processing（安徽省生物医学成像与智能处理重点实验室）

Lubin Gan, Jing Zhang, Heng Zhang, Xin Di, Zhifeng Wang, Wenke Huang, Xiaoyan Sun

备注：: This paper has been withdrawn by the authors due to identified issues in the evaluation protocol in Section Exp. , which may affect the interpretation of the experimental results. The authors are preparing a substantially revised version addressing these issues

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

提出ReconMIL框架，通过潜在空间重建模块适配任务特定特征，并设计双流架构（Mamba全局流+CNN局部流）平衡全局与局部特征，有效抑制背景噪声并定位细粒度诊断区域，在多个基准上超越现有方法。

摘要：

全切片图像（WSI）分析严重依赖于多实例学习（MIL）。尽管近期方法受益于大规模基础模型和高级序列建模以捕获长程依赖，但仍面临两个关键问题。首先，直接使用冻结的、任务无关的特征常因与特定组织学任务的领域差距而导致次优的可分离性。其次，仅依赖全局聚合器可能导致过度平滑，其中稀疏但关键的诊断信号被主导的背景上下文所掩盖。本文提出ReconMIL，一种新颖框架，旨在弥合这一领域差距并平衡全局-局部特征聚合。我们的方法引入潜在空间重建模块，自适应地将通用特征投影到紧凑的任务特定流形中，改善边界划分。为防止信息稀释，我们开发了双流架构，结合基于Mamba的全局流用于上下文先验，以及基于CNN的局部流用于保留细微形态异常。尺度自适应选择机制动态融合这两个流，决定何时依赖整体架构与局部显著性。在多个诊断和生存预测基准上的评估表明，ReconMIL一致优于当前最先进方法，有效定位细粒度诊断区域同时抑制背景噪声。可视化结果证实了模型通过有效平衡全局结构与局部粒度来定位诊断区域的优越能力。

英文摘要：

Whole slide image (WSI) analysis heavily relies on multiple instance learning (MIL). While recent methods benefit from large-scale foundation models and advanced sequence modeling to capture long-range dependencies, they still struggle with two critical issues. First, directly applying frozen, task-agnostic features often leads to suboptimal separability due to the domain gap with specific histological tasks. Second, relying solely on global aggregators can cause over-smoothing, where sparse but critical diagnostic signals are overshadowed by the dominant background context. In this paper, we present ReconMIL, a novel framework designed to bridge this domain gap and balance global-local feature aggregation. Our approach introduces a Latent Space Reconstruction module that adaptively projects generic features into a compact, task-specific manifold, improving boundary delineation. To prevent information dilution, we develop a bi-stream architecture combining a Mamba-based global stream for contextual priors and a CNN-based local stream to preserve subtle morphological anomalies. A scale-adaptive selection mechanism dynamically fuses these two streams, determining when to rely on overall architecture versus local saliency. Evaluations across multiple diagnostic and survival prediction benchmarks show that ReconMIL consistently outperforms current state-of-the-art methods, effectively localizing fine-grained diagnostic regions while suppressing background noise. Visualization results confirm the models superior ability to localize diagnostic regions by effectively balancing global structure and local granularity.

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。