Federated Active Learning Framework for Efficient Annotation Strategy in Skin-lesion Classification

TL;DR

This paper introduces a novel federated active learning framework for skin-lesion classification that reduces annotation effort and preserves privacy, achieving high performance with only half the data in medical image analysis.

Contribution

It presents the first federated active learning framework for medical images, combining ensemble-entropy-based AL with FL to reduce annotation needs while maintaining model accuracy.

Findings

Achieved state-of-the-art performance with only 50% of data.

Outperformed several existing AL methods in federated settings.

Performed comparably to full-data FL in skin-lesion classification.

Abstract

Federated Learning (FL) enables multiple institutes to train models collaboratively without sharing private data. Current FL research focuses on communication efficiency, privacy protection, and personalization and assumes that the data of FL have already been ideally collected. In medical scenarios, however, data annotation demands both expertise and intensive labor, which is a critical problem in FL. Active learning (AL), has shown promising performance in reducing the number of data annotations in medical image analysis. We propose a federated AL (FedAL) framework in which AL is executed periodically and interactively under FL. We exploit a local model in each hospital and a global model acquired from FL to construct an ensemble. We use ensemble-entropy-based AL as an efficient data-annotation strategy in FL. Therefore, our FedAL framework can decrease the amount of annotated data and preserve patient privacy while maintaining the performance of FL. To our knowledge, this is the first FedAL framework applied to medical images. We validated our framework on real-world dermoscopic datasets. Using only 50% of samples, our framework was able to achieve state-of-the-art performance on a skin-lesion classification task. Our framework performed better than several state-of-the-art AL methods under FL and achieved comparable performance to full-data FL.