CohortFinder：一种用于生物医学图像队列数据驱动划分以产生鲁棒机器学习模型的开源工具
机器学习
2023-07-18 v1 计算机视觉与模式识别
摘要
批次效应（BEs）指数据收集中与生物变异无关的系统技术差异，其噪声已被证明会对机器学习（ML）模型的泛化性产生负面影响。在此我们发布 CohortFinder，一种旨在通过数据驱动队列划分缓解 BEs 的开源工具。我们展示 CohortFinder 在下游医学图像处理任务中提升了 ML 模型性能。CohortFinder 可在 cohortfinder.com 免费下载。
引用
@article{arxiv.2307.08673,
title = {CohortFinder: an open-source tool for data-driven partitioning of biomedical image cohorts to yield robust machine learning models},
author = {Fan Fan and Georgia Martinez and Thomas Desilvio and John Shin and Yijiang Chen and Bangchen Wang and Takaya Ozeki and Maxime W. Lafarge and Viktor H. Koelzer and Laura Barisoni and Anant Madabhushi and Satish E. Viswanath and Andrew Janowczyk},
journal= {arXiv preprint arXiv:2307.08673},
year = {2023}
}
备注
26 pages, 9 figures, 4 tables. Abstract was accepted by European Society of Digital and Integrative Pathology (ESDIP), Germany, 2022