FedSAF: A Federated Learning Framework for Enhanced Gastric Cancer Detection and Privacy Preservation

TL;DR

FedSAF tackles the privacy-preserving, data-heterogeneous challenge of gastric cancer detection by integrating model splitting, attentive message passing, and Fisher information-based weighting in a federated setting. The method reduces communication while preserving personalization by sharing only shallow layers and freezing deeper ones, and it dynamically weighs client contributions using AMP and FIM to mitigate non-IID effects. Empirical results on gastric datasets SEED and BOT show superior test accuracy over FedAMP, FedAvg, and FedProx, with strong generalization on FashionMNIST and CIFAR-10, and ablation studies confirm the complementary value of both AMP and FIM. This framework has practical implications for cross-institution collaboration in medical imaging, enabling accurate, privacy-conscious deployment with robust performance across heterogeneous data sources.

Abstract

Gastric cancer is one of the most commonly diagnosed cancers and has a high mortality rate. Due to limited medical resources, developing machine learning models for gastric cancer recognition provides an efficient solution for medical institutions. However, such models typically require large sample sizes for training and testing, which can challenge patient privacy. Federated learning offers an effective alternative by enabling model training across multiple institutions without sharing sensitive patient data. This paper addresses the limited sample size of publicly available gastric cancer data with a modified data processing method. This paper introduces FedSAF, a novel federated learning algorithm designed to improve the performance of existing methods, particularly in non-independent and identically distributed (non-IID) data scenarios. FedSAF incorporates attention-based message passing and the Fisher Information Matrix to enhance model accuracy, while a model splitting function reduces computation and transmission costs. Hyperparameter tuning and ablation studies demonstrate the effectiveness of this new algorithm, showing improvements in test accuracy on gastric cancer datasets, with FedSAF outperforming existing federated learning methods like FedAMP, FedAvg, and FedProx. The framework's robustness and generalization ability were further validated across additional datasets (SEED, BOT, FashionMNIST, and CIFAR-10), achieving high performance in diverse environments.

Figures (9)

• Figure 2: The Structure of FedSAF
• Figure 3: Data Processing
• Figure 4: Averaged and Std Test Accuracy over Rounds for Different Component Combination
• Figure 5: Averaged and Std Test Accuracy over Rounds for Different Distances
• Figure 6: Comparison of Total Transmitted Parameters and Test Accuracy