arXiv 2503.206532025-07-10eess.IVcs.CV

UWarp：一种用于表征扫描仪引起的局部域偏移的全切片图像配准流水线

UWarp: A Whole Slide Image Registration Pipeline to Characterize Scanner-Induced Local Domain Shift

Ummon HealthTech SAS；Centre Georges-François Leclerc（乔治-弗朗索瓦·勒克莱尔中心）；ICMUB laboratory, Université de Bourgogne Europe（勃艮第欧洲大学ICMUB实验室）；University Hospital of Dijon（第戎大学医院）；Hospital de Tortosa Verge de la Cinta (HTVC), Catalan Health Institute (ICS)（托尔托萨维尔赫-德拉辛塔医院 (HTVC)，加泰罗尼亚卫生研究所 (ICS)）；Pere Virgili Institute for Health Research (IISPV)（佩雷·维尔吉利健康研究所 (IISPV)）；Universitat Rovira i Virgili (URV)（罗维拉-威尔吉利大学 (URV)）；Nałęcz Institute of Biocybernetics and Biomedical Engineering Polish Academy of Sciences（波兰科学院纳文奇生物控制论与生物医学工程研究所）；Cairo University（开罗大学）；University of Udine（乌迪内大学）

Antoine Schieb, Bilal Hadjadji, Natalia Fernanda Valderrama, Daniel Tshokola Mweze, Valentin Derangère, Laurent Arnould, Sylvain Ladoire, Alain Lalande, Alessio… 展开作者

Antoine Schieb, Bilal Hadjadji, Natalia Fernanda Valderrama, Daniel Tshokola Mweze, Valentin Derangère, Laurent Arnould, Sylvain Ladoire, Alain Lalande, Alessio Fiorin, Carlos López Pablo, Noèlia Gallardo Borràs, Shrief Abdelazeez, Vincenzo Della Mea, Anna Korzynska, Louis-Oscar Morel, Nathan Vinçon

备注：: preprint

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

提出基于UWarp的全切片图像配准流水线，结合全局仿射变换与局部校正实现高精度对齐，用于表征扫描仪引起的局部域偏移，发现深度学习预测变异与组织密度强相关。

摘要：

组织病理学切片数字化引入了扫描仪引起的域偏移，这可能显著影响基于深度学习方法的计算病理学模型。在现有技术中，这种偏移通常在宏观尺度（切片级或数据集级）进行表征，而非块级，这限制了我们对局部组织特征对深度学习模型准确性的影响的理解。为了应对这一挑战，我们提出了一种基于UWarp的域偏移分析框架，UWarp是一种新颖的配准工具，旨在准确对齐在不同条件下扫描的组织学切片。UWarp采用分层配准方法，结合全局仿射变换与细粒度的局部校正，以实现稳健的组织块对齐。我们使用两个私有数据集CypathLung和BosomShieldBreast评估UWarp，这些数据集包含由多台设备扫描的全切片图像。我们的实验表明，UWarp优于现有的开源配准方法，实现了小于4像素的中位目标配准误差（TRE）（在40倍放大下小于1微米），同时显著减少了计算时间。此外，我们应用UWarp来表征扫描仪引起的局部域偏移对Breast-NEOprAIdict预测的影响，该模型是一种用于乳腺癌病理反应预测的深度学习模型。我们发现预测变异与给定块上的组织密度强相关。我们的研究结果强调了局部域偏移分析的重要性，并表明UWarp可以作为改进计算病理学中模型稳健性和域适应策略的有价值工具。

英文摘要：

Histopathology slide digitization introduces scanner-induced domain shift that can significantly impact computational pathology models based on deep learning methods. In the state-of-the-art, this shift is often characterized at a broad scale (slide-level or dataset-level) but not patch-level, which limits our comprehension of the impact of localized tissue characteristics on the accuracy of the deep learning models. To address this challenge, we present a domain shift analysis framework based on UWarp, a novel registration tool designed to accurately align histological slides scanned under varying conditions. UWarp employs a hierarchical registration approach, combining global affine transformations with fine-grained local corrections to achieve robust tissue patch alignment. We evaluate UWarp using two private datasets, CypathLung and BosomShieldBreast, containing whole slide images scanned by multiple devices. Our experiments demonstrate that UWarp outperforms existing open-source registration methods, achieving a median target registration error (TRE) of less than 4 pixels (<1 micrometer at 40x magnification) while significantly reducing computational time. Additionally, we apply UWarp to characterize scanner-induced local domain shift in the predictions of Breast-NEOprAIdict, a deep learning model for breast cancer pathological response prediction. We find that prediction variability is strongly correlated with tissue density on a given patch. Our findings highlight the importance of localized domain shift analysis and suggest that UWarp can serve as a valuable tool for improving model robustness and domain adaptation strategies in computational pathology.

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。

AI总结：

摘要：

英文摘要：

相关论文