arXivDaily arXiv每日学术速递 周一至周五更新
​通知:联邦节假日调休:哥伦布日 / 原住民日(Columbus Day / Indigenous Peoples' Day),今日(13日)无更新。
arXiv 2412.14545cs.CVcs.AIcs.LG

基于点Transformer和联邦学习的H&E染色全切片图像预测乳腺癌HER2状态总结

Summary of Point Transformer with Federated Learning for Predicting Breast Cancer HER2 Status from Hematoxylin and Eosin-Stained Whole Slide Images

Towson University(陶森大学)

Kamorudeen A. Amuda, Almustapha A. Wakili

AI总结:

本研究提出基于联邦学习和点Transformer的方法,结合动态标签分布、辅助分类器与最远余弦采样,从H&E染色全切片图像预测乳腺癌HER2状态,在多站点数据上表现卓越且泛化性强。

摘要:

本研究引入基于联邦学习的方法,从H&E染色全切片图像(WSIs)预测HER2状态,降低成本并加速治疗决策。针对多站点数据集中的标签不平衡和特征表示挑战,提出点Transformer,结合动态标签分布、辅助分类器和最远余弦采样。大量实验表明,在四个站点(2687张WSIs)上达到SOTA性能,对两个未见站点(229张WSIs)具有强泛化能力。

英文摘要:

This study introduces a federated learning-based approach to predict HER2 status from hematoxylin and eosin (HE)-stained whole slide images (WSIs), reducing costs and speeding up treatment decisions. To address label imbalance and feature representation challenges in multisite datasets, a point transformer is proposed, incorporating dynamic label distribution, an auxiliary classifier, and farthest cosine sampling. Extensive experiments demonstrate state-of-the-art performance across four sites (2687 WSIs) and strong generalization to two unseen sites (229 WSIs).

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。

↑