SelfFed: Self-Supervised Federated Learning for Data Heterogeneity and Label Scarcity in Medical Images

TL;DR

SelfFed is a novel federated learning framework that effectively addresses data heterogeneity and label scarcity in medical imaging by combining pre-training, contrastive learning, and a new aggregation strategy, leading to improved performance on challenging datasets.

Contribution

The paper introduces SelfFed, a self-supervised federated learning approach that tackles data heterogeneity and label scarcity in medical images through a two-phase process with innovative strategies.

Findings

Achieves up to 8.8% and 4.1% improvements on Retina and COVID-FL datasets.

Performs well even with only 10% labeled data.

Outperforms existing federated learning baselines.

Abstract

Self-supervised learning in the federated learning paradigm has been gaining a lot of interest both in industry and research due to the collaborative learning capability on unlabeled yet isolated data. However, self-supervised based federated learning strategies suffer from performance degradation due to label scarcity and diverse data distributions, i.e., data heterogeneity. In this paper, we propose the SelfFed framework for medical images to overcome data heterogeneity and label scarcity issues. The first phase of the SelfFed framework helps to overcome the data heterogeneity issue by leveraging the pre-training paradigm that performs augmentative modeling using Swin Transformer-based encoder in a decentralized manner. The label scarcity issue is addressed by fine-tuning paradigm that introduces a contrastive network and a novel aggregation strategy. We perform our experimental analysis on publicly available medical imaging datasets to show that SelfFed performs better when compared to existing baselines and works. Our method achieves a maximum improvement of 8.8% and 4.1% on Retina and COVID-FL datasets on non-IID datasets. Further, our proposed method outperforms existing baselines even when trained on a few (10%) labeled instances.