arXiv 2510.065922025-10-08cs.CV

自适应染色归一化用于跨域医学组织学

Adaptive Stain Normalization for Cross-Domain Medical Histology

Dept. of Biomedical Engineering, Johns Hopkins University, Baltimore, MD, USA（生物医学工程系，约翰霍普金斯大学，巴尔的摩，MD，USA）；Johns Hopkins Malaria Research Institute, W. Harry Feinstone Department of Molecular Microbiology and Immunology, Bloomberg School of Public Health, Baltimore, MD, USA（约翰霍普金斯疟疾研究 institute，W. Harry Feinstone 分子微生物学与免疫学系，布隆伯格公共卫生学院，巴尔的摩，MD，USA）；Dept. of Medicine, Johns Hopkins University School of Medicine, Baltimore, MD, USA（医学系，约翰霍普金斯大学医学院，巴尔的摩，MD，USA）；Dept. of Electrical and Systems Engineering, University of Pennsylvania, Philadelphia, PA, USA（电气与系统工程系，宾夕法尼亚大学，费城，PA，USA）

Tianyue Xu, Yanlin Wu, Abhai K. Tripathi, Matthew M. Ippolito, Benjamin D. Haeffele

备注：: Accepted to the 28th International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI 2025)
DOI：: 10.1007/978-3-032-04981-0_3

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

针对病理图像染色差异导致的域偏移问题，提出基于比尔-朗伯定律和非负矩阵分解算法展开的可训练颜色归一化模型，可集成于任意骨干网络，在跨域检测与分类任务上优于现有方法。

摘要：

深度学习的进展彻底改变了自动化数字病理学分析。然而，染色方案和成像条件的差异会引入显著的颜色变异性。在深度学习中，当模型部署在与训练数据不同条件下获取的数据上时，这种颜色不一致性通常会降低性能，这一挑战被称为域偏移。许多现有方法试图通过颜色归一化来解决这个问题，但存在几个明显的缺点，例如引入伪影或需要仔细选择用于染色映射的模板图像。为了解决这些局限性，我们提出了一种可训练的颜色归一化模型，该模型可以与任何骨干网络集成，用于下游任务，如目标检测和分类。基于比尔-朗伯定律的成像过程物理原理，我们的模型架构通过非负矩阵分解（NMF）模型的算法展开推导而来，以从原始病理图像中提取染色不变的结构信息，该信息作为进一步处理的输入。在实验方面，我们在公开可用的病理数据集和内部整理的疟疾血涂片集合上评估了该方法，用于跨域目标检测和分类，我们的方法优于许多最先进的染色归一化方法。我们的代码可在 https://github.com/xutianyue/BeerLaNet 获取。

英文摘要：

Deep learning advances have revolutionized automated digital pathology analysis. However, differences in staining protocols and imaging conditions can introduce significant color variability. In deep learning, such color inconsistency often reduces performance when deploying models on data acquired under different conditions from the training data, a challenge known as domain shift. Many existing methods attempt to address this problem via color normalization but suffer from several notable drawbacks such as introducing artifacts or requiring careful choice of a template image for stain mapping. To address these limitations, we propose a trainable color normalization model that can be integrated with any backbone network for downstream tasks such as object detection and classification. Based on the physics of the imaging process per the Beer-Lambert law, our model architecture is derived via algorithmic unrolling of a nonnegative matrix factorization (NMF) model to extract stain-invariant structural information from the original pathology images, which serves as input for further processing. Experimentally, we evaluate the method on publicly available pathology datasets and an internally curated collection of malaria blood smears for cross-domain object detection and classification, where our method outperforms many state-of-the-art stain normalization methods. Our code is available at https://github.com/xutianyue/BeerLaNet.

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。