Federated Fine-tuning of SAM-Med3D for MRI-based Dementia Classification

2025/08/29 by Mouheb, Kaouther, Elbatel, Marawan, Papma, Janne +8
#Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences

paper · doi:10.48550/arxiv.2508.21458

Abstract

While foundation models (FMs) offer strong potential for AI-based dementia diagnosis, their integration into federated learning (FL) systems remains underexplored. In this benchmarking study, we systematically evaluate the impact of key design choices: classification head architecture, fine-tuning strategy, and aggregation method, on the performance and efficiency of federated FM tuning using brain MRI data. Using a large multi-cohort dataset, we find that the architecture of the classification head substantially influences performance, freezing the FM encoder achieves comparable results to full fine-tuning, and advanced aggregation methods outperform standard federated averaging. Our results offer practical insights for deploying FMs in decentralized clinical settings and highlight trade-offs that should guide future method development.

Citations

• Rate-My-LoRA: Efficient and Adaptive Federated Model Tuning for Cardiac MRI Segmentation
• Probing the Efficacy of Federated Parameter-Efficient Fine-Tuning of Vision Transformers for Medical Image Classification
• SAM-Med3D: Towards General-purpose Segmentation Models for Volumetric Medical Images
• Parameter-Efficient Fine-Tuning for Medical Image Analysis: The Missed Opportunity
• DINOv2: Learning Robust Visual Features without Supervision
• Fair Federated Medical Image Segmentation via Client Contribution Estimation
• MONAI: An open-source framework for deep learning in healthcare
• Self-Supervised Pre-Training of Swin Transformers for 3D Medical Image Analysis
• On the Opportunities and Risks of Foundation Models
• LoRA: Low-Rank Adaptation of Large Language Models
• Open Access Series of Imaging Studies (OASIS): Cross-sectional MRI Data in Young, Middle Aged, Nondemented, and Demented Older Adults
• Evaluating General Purpose Vision Foundation Models for Medical Image Analysis: An Experimental Study of DINOv2 on Radiology Benchmarks

Related

• Decoding the Alzheimer's Continuum: Interpretable Multi-Gate Routing for Diagnosis and Transition Prediction
• Federated Learning for Predicting Mild Cognitive Impairment to Dementia Conversion
• A Multi-Frequency Self-Supervised Fusion Model for EEG-Based Dementia Classification
• UG-FedDA: Uncertainty-Guided Federated Domain Adaptation for Multi-Center Alzheimer's Disease Detection
• Dementia Etiology Diagnosis via Collaborative Meta Knowledge Enhancement