FedST: Federated Style Transfer Learning for Non-IID Image Segmentation
Boyuan Ma, Xiang Yin, Jing Tan, Yongfeng Chen, Haiyou Huang, Hao Wang, Weihua Xue, Xiaojuan Ban
Abstract
Federated learning collaboratively trains machine learning models among different clients while keeping data privacy and has become the mainstream for breaking data silos. However, the non-independently and identically distribution (i.e., Non-IID) characteristic of different image domains among different clients reduces the benefits of federated learning and has become a bottleneck problem restricting the accuracy and generalization of federated models. In this work, we propose a novel federated image segmentation method based on style transfer, FedST, by using a denoising diffusion probabilistic model to achieve feature disentanglement and image synthesis of cross-domain image data between multiple clients. Thus it can share style features among clients while protecting structure features of image data, which effectively alleviates the influence of the Non-IID phenomenon. Experiments prove that our method achieves superior segmentation performance compared to state-of-art methods among four different Non-IID datasets in objective and subjective assessment. The code is available at https://github.com/YoferChen/FedST.
BibTeX
@article{Ma_Yin_Tan_Chen_Huang_Wang_Xue_Ban_2024, title={FedST: Federated Style Transfer Learning for Non-IID Image Segmentation}, volume={38}, url={https://ojs.aaai.org/index.php/AAAI/article/view/28199}, DOI={10.1609/aaai.v38i5.28199}, abstractNote={Federated learning collaboratively trains machine learning models among different clients while keeping data privacy and has become the mainstream for breaking data silos. However, the non-independently and identically distribution (i.e., Non-IID) characteristic of different image domains among different clients reduces the benefits of federated learning and has become a bottleneck problem restricting the accuracy and generalization of federated models. In this work, we propose a novel federated image segmentation method based on style transfer, FedST, by using a denoising diffusion probabilistic model to achieve feature disentanglement and image synthesis of cross-domain image data between multiple clients. Thus it can share style features among clients while protecting structure features of image data, which effectively alleviates the influence of the Non-IID phenomenon. Experiments prove that our method achieves superior segmentation performance compared to state-of-art methods among four different Non-IID datasets in objective and subjective assessment. The code is available at https://github.com/YoferChen/FedST.}, number={5}, journal={Proceedings of the AAAI Conference on Artificial Intelligence}, author={Ma, Boyuan and Yin, Xiang and Tan, Jing and Chen, Yongfeng and Huang, Haiyou and Wang, Hao and Xue, Weihua and Ban, Xiaojuan}, year={2024}, month={Mar.}, pages={4053-4061} }