What does this research mean for the field?

The FedFB federated learning framework achieves 91.07% accuracy in fetal brain ultrasound classification under non-IID conditions while reducing communication overhead by over 85% compared to traditional methods. Novelty: ClaimNovelty.METHODOLOGICAL. Consensus alignment: ConsensusAlignment.NEUTRAL.

What question did this study set out to answer?

The aim is to develop a federated learning framework that improves model performance on non-IID datasets while reducing communication overhead.

March 24, 2026Open Access

A New Communication Efficient Federated Online Distillation for Non-IID Data: Application to Fetal Brain Ultrasound Plane Classification

Key Result

The FedFB federated learning framework achieved 91.07% accuracy in fetal brain ultrasound classification and reduced communication overhead by over 85% compared to traditional methods.

Key Points

The aim is to develop a federated learning framework that improves model performance on non-IID datasets while reducing communication overhead.
Developed the FedFB framework integrating online ensemble knowledge distillation.
Utilized multiple locally trained teacher models for knowledge distillation into a lightweight student network.
Incorporated an auxiliary attention–convolution module to enhance model performance.
Leveraged a small public dataset to facilitate knowledge transfer without disclosing private data.
Achieved 91.07% accuracy, with high precision, recall, and F1-score metrics under non-IID conditions.
Reduced communication overhead by more than 85% compared to traditional federated learning methods.
Demonstrated robustness across multiple medical imaging datasets, including MRI and X-rays.

Structured PICO

Population

Fetal brain ultrasound datasets, brain tumor MRI, and chest X-ray datasets

Intervention

FedFB (communication-efficient federated learning framework integrating online ensemble knowledge distillation)

Comparator

Traditional federated learning (FL) methods

Outcome

Classification accuracy, precision, recall, F1-score, and communication overhead

The FedFB framework enables privacy-aware, communication-efficient federated learning for medical imaging, maintaining high classification performance under non-IID data conditions.

Main Result

Absolute Event Rate: 0% vs 0%

Abstract

Machine learning has achieved notable success in medical imaging; however, its reliability remains constrained by the quality and quantity of training data, particularly in specialized domains such as fetal brain ultrasound. Federated learning (FL) offers a collaborative alternative by enabling multiple institutions to train shared models without exposing sensitive data. Yet, its effectiveness often degrades under non-independent and identically distributed (non-IID) data and its communication demands can be prohibitive. To address these challenges, we propose FedFB, a communication-efficient federated learning framework that integrates online ensemble knowledge distillation for privacy-aware collaborative learning. In FedFB, multiple teacher models trained locally on private datasets collectively distill their knowledge into a lightweight dual-branch student network enhanced with an auxiliary attention–convolution module, without increasing model complexity. The distillation process leverages a small auxiliary public dataset for soft-label transfer, reducing data exposure by avoiding the exchange of raw private data or model parameters, while relying on shared supervisory signals for knowledge transfer. Experimental results on fetal brain ultrasound datasets demonstrate that FedFB achieves 91.07% accuracy, 91.28% precision, 91.03% recall, and 91.15% F1-score under the non-IID conditions, while reducing communication overhead by more than 85% compared to traditional FL methods. Furthermore, the framework’s robustness was validated on brain tumor MRI and chest X-ray datasets, confirming its generalization capability across distinct medical imaging modalities.