Federated learning (FL) has emerged as a promising paradigm for training segmentation models on decentralized medical data, owing to its privacy-preserving property. However, existing research overlooks the prevalent annotation noise encountered in real-world medical datasets, which limits the performance ceilings of FL. In this paper, we, for the first time, identify and tackle this problem. For problem formulation, we propose a contour evolution for modeling non-independent and identically distributed (Non-IID) noise across pixels within each client and then extend it to the case of multi-source data to form a heterogeneous noise model (\textit{i.e.}, Non-IID annotation noise across clients). For robust learning from annotations with such two-level Non-IID noise, we emphasize the importance of data quality in model aggregation, allowing high-quality clients to have a greater impact on FL. To achieve this, we propose \textbf{Fed}erated learning with \textbf{A}nnotation qu\textbf{A}lity-aware \textbf{A}ggregat\textbf{I}on, named \textbf{FedA$^3$I}, by introducing a quality factor based on client-wise noise estimation. Specifically, noise estimation at each client is accomplished through the Gaussian mixture model and then incorporated into model aggregation in a layer-wise manner to up-weight high-quality clients. Extensive experiments on two real-world medical image segmentation datasets demonstrate the superior performance of FedA$^3$I against the state-of-the-art approaches in dealing with cross-client annotation noise. The code is available at \color{blue}{https://github.com/wnn2000/FedAAAI}.

翻译：联邦学习（FL）作为一种有前景的范式，凭借其隐私保护特性，能够利用分散的医疗数据训练分割模型。然而，现有研究忽略了真实医疗数据集中普遍存在的标注噪声问题，这限制了FL的性能上限。本文首次识别并解决了该问题。在问题建模方面，我们提出了一种轮廓演化方法，用于模拟客户端内像素间的非独立同分布（Non-IID）噪声，并将其扩展至多源数据场景，形成异构噪声模型（即客户端间的Non-IID标注噪声）。针对此类双层Non-IID噪声下的鲁棒学习，我们强调数据质量在模型聚合中的重要性，使高质量客户端对FL产生更大影响。为此，我们提出带有标注质量感知聚合的联邦学习（简称FedA$^3$I），通过引入基于客户端噪声估计的质量因子实现。具体而言，每个客户端的噪声估计通过高斯混合模型完成，随后以逐层方式融入模型聚合，以提升高质量客户端的权重。在两个真实医学图像分割数据集上的大量实验表明，FedA$^3$I在处理跨客户端标注噪声方面优于现有最先进方法。代码已开源在\color{blue}{https://github.com/wnn2000/FedAAAI}。