对比三类联邦学习方法在脑肿瘤分割上的表现，验证其隐私保护下的有效性。
Federated brain tumor segmentation: an extensive benchmark
- 按全局、个性化、混合三类设计联邦学习基准测试
- 标准FedAvg已表现优异，部分方法可微调提升性能
- 揭示数据分布对模型偏差的影响，适合医疗图像协作研究者
近年来，联邦学习因其在保护隐私的前提下聚合多中心数据的能力，在医学图像分析领域引发广泛关注。已有大量联邦训练方案被提出，我们将其分为全局（一个最终模型）、个性化（每机构一个模型）和混合（每集群一个模型）三类。然而，这些方法在最新发布的Federated Brain Tumor Segmentation 2022数据集上的适用性尚未探索。本文针对该任务构建了涵盖三类方法的全面基准测试。尽管标准FedAvg已表现良好，但各类别中部分方法仍能带来轻微性能提升，并可能缓解最终模型对联邦主导数据分布的偏差。此外，我们通过独立同分布（IID）设置和有限数据设置，进一步分析了联邦学习在此任务中的行为特征。
原文摘要 · Abstract (English)
Recently, federated learning has raised increasing interest in the medical image analysis field due to its ability to aggregate multi-center data with privacy-preserving properties. A large amount of federated training schemes have been published, which we categorize into global (one final model), personalized (one model per institution) or hybrid (one model per cluster of institutions) methods. However, their applicability on the recently published Federated Brain Tumor Segmentation 2022 dataset has not been explored yet. We propose an extensive benchmark of federated learning algorithms from all three classes on this task. While standard FedAvg already performs very well, we show that some methods from each category can bring a slight performance improvement and potentially limit the final model(s) bias toward the predominant data distribution of the federation. Moreover, we provide a deeper understanding of the behaviour of federated learning on this task through alternative ways of distributing the pooled dataset among institutions, namely an Independent and Identical Distributed (IID) setup, and a limited data setup.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。