跳到正文
arXiv cs.AI· Tao Lian, Jose L. G\'omez, Antonio M. L\'opez·· 10 小时前AI 评分22

FedS2R:面向自动驾驶合成到真实语义分割的一次性联邦域泛化框架

FedS2R: One-Shot Federated Domain Generalization for Synthetic-to-Real Semantic Segmentation in Autonomous Driving

AI 导读

FedS2R 是首个面向自动驾驶合成到真实语义分割的一次性联邦域泛化框架,由不一致性驱动的数据增强策略和多客户端知识蒸馏与特征融合两部分组成。在 Cityscapes、BDD100K、Mapillary、IDD 和 ACDC 五个真实数据集上,其全局模型显著优于各客户端单独模型,仅比同时访问全部客户端数据训练的模型低 2 mIoU。

正文 · 原文

View PDF HTML (experimental)

Abstract:Federated domain generalization has shown promising progress in image classification by enabling collaborative training across multiple clients without sharing raw data. However, its potential in the semantic segmentation of autonomous driving remains underexplored. In this paper, we propose FedS2R, the first one-shot federated domain generalization framework for synthetic-to-real semantic segmentation in autonomous driving. FedS2R comprises two components: an inconsistency-driven data augmentation strategy that generates images for unstable classes, and a multi-client knowledge distillation scheme with feature fusion that distills a global model from multiple client models. Experiments on five real-world datasets, Cityscapes, BDD100K, Mapillary, IDD, and ACDC, show that the global model significantly outperforms individual client models and is only 2 mIoU points behind the model trained with simultaneous access to all client data. These results demonstrate the effectiveness of FedS2R in synthetic-to-real semantic segmentation for autonomous driving under federated learning
Comments: Accepted by IEEE Intelligent Vehicles Symposium (IV) 2026
Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
Cite as: arXiv:2507.19881 [cs.CV]
  (or arXiv:2507.19881v4 [cs.CV] for this version)
  https://doi.org/10.48550/arXiv.2507.19881

arXiv-issued DOI via DataCite

Submission history

From: Tao Lian [view email]
[v1] Sat, 26 Jul 2025 09:24:00 UTC (5,730 KB)
[v2] Mon, 1 Jun 2026 14:48:08 UTC (5,839 KB)
[v3] Tue, 21 Jul 2026 10:06:33 UTC (5,839 KB)
[v4] Wed, 18 Nov 2026 10:41:29 UTC (5,838 KB)

来源:arXiv cs.AI · arxiv.org