Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

MH-pFLGB: Model Heterogeneous personalized Federated Learning via Global Bypass for Medical Image Analysis

Luyuan Xie, Manqing Lin, ChenMing Xu, Tianyu Luan, Zhipeng Zeng, Wenjun Qian, Cong Li, Yuejian Fang, Qingni Shen, Zhonghai Wu

TL;DR

MH-pFLGB addresses federated learning for medical imaging under statistical and system heterogeneity while reducing reliance on public datasets. It introduces a lightweight global bypass model with a body/head design and a feature weighted fusion that integrates global and local representations across heterogeneous clients. Across medical image classification at multiple resolutions, classification under different label distributions, and segmentation tasks, MH-pFLGB achieves state-of-the-art performance, outperforming methods that rely on soft predictions from public data. The approach offers privacy-preserving, computation-efficient collaboration with broad applicability to medical imaging tasks such as classification and segmentation.

Abstract

In the evolving application of medical artificial intelligence, federated learning is notable for its ability to protect training data privacy. Federated learning facilitates collaborative model development without the need to share local data from healthcare institutions. Yet, the statistical and system heterogeneity among these institutions poses substantial challenges, which affects the effectiveness of federated learning and hampers the exchange of information between clients. To address these issues, we introduce a novel approach, MH-pFLGB, which employs a global bypass strategy to mitigate the reliance on public datasets and navigate the complexities of non-IID data distributions. Our method enhances traditional federated learning by integrating a global bypass model, which would share the information among the clients, but also serves as part of the network to enhance the performance on each client. Additionally, MH-pFLGB provides a feature fusion module to better combine the local and global features. We validate \model{}'s effectiveness and adaptability through extensive testing on different medical tasks, demonstrating superior performance compared to existing state-of-the-art methods.

MH-pFLGB: Model Heterogeneous personalized Federated Learning via Global Bypass for Medical Image Analysis

TL;DR

MH-pFLGB addresses federated learning for medical imaging under statistical and system heterogeneity while reducing reliance on public datasets. It introduces a lightweight global bypass model with a body/head design and a feature weighted fusion that integrates global and local representations across heterogeneous clients. Across medical image classification at multiple resolutions, classification under different label distributions, and segmentation tasks, MH-pFLGB achieves state-of-the-art performance, outperforming methods that rely on soft predictions from public data. The approach offers privacy-preserving, computation-efficient collaboration with broad applicability to medical imaging tasks such as classification and segmentation.

Abstract

In the evolving application of medical artificial intelligence, federated learning is notable for its ability to protect training data privacy. Federated learning facilitates collaborative model development without the need to share local data from healthcare institutions. Yet, the statistical and system heterogeneity among these institutions poses substantial challenges, which affects the effectiveness of federated learning and hampers the exchange of information between clients. To address these issues, we introduce a novel approach, MH-pFLGB, which employs a global bypass strategy to mitigate the reliance on public datasets and navigate the complexities of non-IID data distributions. Our method enhances traditional federated learning by integrating a global bypass model, which would share the information among the clients, but also serves as part of the network to enhance the performance on each client. Additionally, MH-pFLGB provides a feature fusion module to better combine the local and global features. We validate \model{}'s effectiveness and adaptability through extensive testing on different medical tasks, demonstrating superior performance compared to existing state-of-the-art methods.
Paper Structure (14 sections, 6 equations, 2 figures, 5 tables)

This paper contains 14 sections, 6 equations, 2 figures, 5 tables.

Table of Contents

  1. Introduction
  2. Methods
  3. Pipeline
  4. Features Weighted Fusion
  5. Experiments Setup
  6. Results and Discussion
  7. Medical Image Classification (Different Resolutions)
  8. Medical Image Classification (Different Label Distributions)
  9. Medical Image Segmentation
  10. Ablation Experiments
  11. GFLOPS and Parameters
  12. Conclusion
  13. Acknowledgements.
  14. Disclosure of Interests.

Figures (2)

  • Figure 1: (a) Overview of our proposed MH-pFLGB framework. Each training process consists of 3 steps. From a to c: a. Local model training. b. Global bypass model training. c. Upload, aggregation, and download. (b) Features weighted fusion. More details can be found in Section 2.1 and Section 2.2.
  • Figure 2: Visualized comparison of Federated Learning in medical image segmentation. We randomly select four samples from different clients to form the visualization. (a-j) Segmentation results by a model trained with FedAVG, SCAFFOLD, FedProx, Ditto, APFL, LG-FedAvg, FedRep, FedSM, LC-Fed, and our method MH-pFLGB; (k) Ground truths (denoted as ‘GT’).