Research map / Federated learning and privacy / Personalized and heterogeneous FL
FedSeg: Class-Heterogeneous Federated Learning for Semantic Segmentation
Jiaxu Miao, Zongxin Yang, Leilei Fan, Yi Yang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Add background-classes aggregation cross-entropy (ablation) | FedAvg semantic segmentation · Cityscapes non-IID1 | mIoU: 10.40 -> 45.08 | Improved |
| Add background-classes aggregation cross-entropy (ablation) | FedAvg semantic segmentation · CamVID non-IID1 | mIoU: 19.06 -> 58.38 | Improved |
| Add background-classes aggregation cross-entropy (ablation) | FedAvg semantic segmentation · PascalVOC | mIoU: 8.56 -> 32.28 | Improved |
| Add local-to-global pixel contrastive loss (ablation) | FedAvg + Lbackce · Cityscapes non-IID1 | mIoU: 45.08 -> 50.24 | Improved |
| Add local-to-global pixel contrastive loss (ablation) | FedAvg + Lbackce · CamVID non-IID1 | mIoU: 58.38 -> 63.50 | Improved |
| Add local-to-global pixel contrastive loss (ablation) | FedAvg + Lbackce · PascalVOC | mIoU: 32.28 -> 32.20 | Hurt |
| Add local-to-global pixel contrastive loss (ablation) | FedAvg + Lbackce · ADE20k | mIoU: 8.31 -> 8.64 | Improved |
| Replace local-to-local pixel contrastive learning with local-to-global contrastive learning (ablation) | FedSeg · Cityscapes non-IID1 | mIoU: 47.91% -> 50.24% | Improved |
| Use pixel-to-region rather than pixel-to-pixel contrastive sampling (ablation) | FedSeg · Cityscapes non-IID1 | mIoU: 49.95% -> 50.24% | No clear effect |
| Increase the number of local epochs per round (ablation) | FedSeg · Cityscapes non-IID1 and CamVID non-IID1 | mIoU | Hurt |
Later papers that compared with this method
1 later accepted paper reports FedSeg as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- Capturing Spectral and Spatial Patterns for Federated Remote Sensing Segmentation (ECCV 2026 · better on 5 of 5 shared results)
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Federated Black-Box Adaptation for Semantic Segmentation (NeurIPS 2024 · similarity 0.89)
- CLIP-Guided Federated Learning on Heterogeneity and Long-Tailed Data (AAAI 2024 · similarity 0.88)
- FedST: Federated Style Transfer Learning for Non-IID Image Segmentation (AAAI 2024 · similarity 0.87)
- FedQS: Optimizing Gradient and Model Aggregation for Semi-Asynchronous Federated Learning (NeurIPS 2025 · similarity 0.86)
- FedDAT: An Approach for Foundation Model Finetuning in Multi-Modal Heterogeneous Federated Learning (AAAI 2024 · similarity 0.86)
- FedBiP: Heterogeneous One-Shot Federated Learning with Personalized Latent Diffusion Models (CVPR 2025 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
