发表机构
Southern Medical University; Huazhong University of Science and Technology; Hubei University of Technology; Tongji Hospital; Hubei Maternal and Child Health Hospital（南方医科大学; 华中科技大学; 湖北工业大学; 同济医院; 湖北省妇幼保健院）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本文提出StainPresetNet框架，解决染色归一化方法的颜色映射不准确、效率低、方向刚性等问题，在病理数据集上验证其精度高、泛化能力强且计算开销低90%的优势。
AI 中文摘要
染色归一化可降低染色方案和成像条件差异导致的颜色变化，从而提升计算机辅助诊断系统的性能。传统方法通过逐像素变换从单个或有限参考图像推导映射关系，具有风格灵活性，但存在颜色映射提取不准确的问题。现有基于深度学习的方法虽能通过复杂神经网络实现数据集范围内的准确颜色映射，但面临计算效率低、伪影产生、归一化方向固定（方向改变需重新训练模型）等挑战。为解决这些局限，本文提出StainPresetNet——一种结合结构保留与数据集级颜色映射同时保持计算效率的新型框架。该方法在预设参考图像的引导下实现逐像素归一化，无需重新训练即可实现多方向适应性。在细胞病理学和组织病理学数据集上的评估表明，StainPresetNet相比传统方法实现了更优的颜色映射精度，有效提升了诊断任务中分类器的泛化能力，且与现有深度学习方法相比计算开销降低了90%。所提出的预设引导机制可通过简单替换参考图像灵活调整归一化方向，克服了当前基于深度学习解决方案的方向刚性问题。
英文摘要
Stain normalization reduces color variations caused by variations in staining protocols and imaging conditions, thereby enhancing computer-aided diagnostic system performance. Traditional methods derive mapping relationships from individual or limited reference images through pixel-wise transformation, offering style flexibility but suffering from inaccurate color mapping extraction. While existing deep-learning-based approaches achieve accurate dataset-wide color mapping through complex neural networks, they face challenges including computational inefficiency, artifact generation, and fixed normalization directions requiring model retraining for directional changes. To address these limitations, we propose StainPresetNet - a novel framework that combines structural preservation with dataset-level color mapping while maintaining computational efficiency. Our method implements pixel-wise normalization guided by preset reference images, enabling multi-directional adaptability without retraining. Evaluations on cytopathology and histopathology datasets demonstrate that StainPresetNet achieves superior color mapping accuracy compared to conventional methods, effectively improves classifier generalization in diagnostic tasks, and reduces computational overhead by 90\% versus existing deep learning approaches. The proposed preset-guided mechanism facilitates flexible adjustment of normalization directions through simple reference image replacement, overcoming the directional rigidity of current deep-learning-based solutions.
