Cross-domain Federated Object Detection

TL;DR

This paper introduces FedOD, a federated learning framework for cross-domain object detection that combines global knowledge and personalized models to improve detection performance across diverse client data distributions.

Contribution

The paper proposes FedOD, a novel federated object detection framework using multi-teacher distillation and personalized fine-tuning for cross-domain scenarios.

Findings

FedOD improves detection accuracy across clients with different data distributions.

The framework effectively combines global and local models through weighted ensemble inference.

Experimental results validate the effectiveness of FedOD on a new cross-domain autonomous driving dataset.

Abstract

Detection models trained by one party (including server) may face severe performance degradation when distributed to other users (clients). Federated learning can enable multi-party collaborative learning without leaking client data. In this paper, we focus on a special cross-domain scenario in which the server has large-scale labeled data and multiple clients only have a small amount of labeled data; meanwhile, there exist differences in data distributions among the clients. In this case, traditional federated learning methods can't help a client learn both the global knowledge of all participants and its own unique knowledge. To make up for this limitation, we propose a cross-domain federated object detection framework, named FedOD. The proposed framework first performs the federated training to obtain a public global aggregated model through multi-teacher distillation, and sends the aggregated model back to each client for fine-tuning its personalized local model. After a few rounds of communication, on each client we can perform weighted ensemble inference on the public global model and the personalized local model. We establish a federated object detection dataset which has significant background differences and instance differences based on multiple public autonomous driving datasets, and then conduct extensive experiments on the dataset. The experimental results validate the effectiveness of the proposed method.