Summary of Pflfe: Cross-silo Personalized Federated Learning Via Feature Enhancement on Medical Image Segmentation, by Luyuan Xie et al.
pFLFE: Cross-silo Personalized Federated Learning via Feature Enhancement on Medical Image Segmentation
by Luyuan Xie, Manqing Lin, Siyuan Liu, ChenMing Xu, Tianyu Luan, Cong Li, Yuejian Fang, Qingni Shen, Zhonghai Wu
First submitted to arxiv on: 29 Jun 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This paper proposes a new framework for personalized cross-silo federated learning (FL) in medical image segmentation. The proposed method, Personalized Federated Learning via Feature Enhancement (pFLFE), aims to mitigate client drift and improve performance by enhancing features and utilizing segmentation masks. The framework consists of two stages: feature enhancement and supervised learning. Experiments on three medical segmentation tasks show that pFLFE outperforms state-of-the-art methods. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Medical image segmentation is a crucial task in healthcare, but it’s challenging due to limited data and privacy concerns. Researchers have been working on personalized cross-silo federated learning (FL) to overcome these challenges. However, existing methods often fail because of client drift, which causes inconsistent performance and slow training. This paper proposes a new way to do FL that solves this problem. It’s called Personalized Federated Learning via Feature Enhancement (pFLFE). The method has two parts: making the features better and using those features for learning from masks. The authors also came up with a new way of training that requires fewer communication rounds without sacrificing quality, even with limited resources. They tested it on three tasks and showed that pFLFE works better than what’s currently available. |
Keywords
» Artificial intelligence » Federated learning » Image segmentation » Supervised
