Federated Learning Method for Preserving Privacy in Face Recognition System

TL;DR

Experimental findings based on the CelebA datasets reveal that employing federated learning in both supervised and unsupervised face recognition systems offers dual benefits, particularly in terms of the balance between privacy and accuracy.

Abstract

The state-of-the-art face recognition systems are typically trained on a single computer, utilizing extensive image datasets collected from various number of users. However, these datasets often contain sensitive personal information that users may hesitate to disclose. To address potential privacy concerns, we explore the application of federated learning, both with and without secure aggregators, in the context of both supervised and unsupervised face recognition systems. Federated learning facilitates the training of a shared model without necessitating the sharing of individual private data, achieving this by training models on decentralized edge devices housing the data. In our proposed system, each edge device independently trains its own model, which is subsequently transmitted either to a secure aggregator or directly to the central server. To introduce diverse data without the need for data transmission, we employ generative adversarial networks to generate imposter data at the edge. Following this, the secure aggregator or central server combines these individual models to construct a global model, which is then relayed back to the edge devices. Experimental findings based on the CelebA datasets reveal that employing federated learning in both supervised and unsupervised face recognition systems offers dual benefits. Firstly, it safeguards privacy since the original data remains on the edge devices. Secondly, the experimental results demonstrate that the aggregated model yields nearly identical performance compared to the individual models, particularly when the federated model does not utilize a secure aggregator. Hence, our results shed light on the practical challenges associated with privacy-preserving face image training, particularly in terms of the balance between privacy and accuracy.

Figures (7)

• Figure 1: The proposed face recognition system incorporates federated learning. Through the implementation of a secure aggregator, we empower a collective of inherently untrusting devices to collaborate and calculate an aggregate value without disclosing their individual private data.
• Figure 2: Histograms depicting the Equal Error Rate (EER) across 1000 devices are presented for the comparison between individual and federated models in the supervised systems. Notably, this evaluation focuses on models that do not utilize a secure aggregator (SA).
• Figure 3: Histograms illustrating the Equal Error Rate (EER) distribution across 1000 devices are provided for a comparison between individual and federated models in the supervised system. This analysis specifically considers models that incorporate a secure aggregator (SA).
• Figure 4: The box plot depicts the distribution of Equal Error Rates (EER) for both supervised individual and federated models across 1000 devices. The analysis considers scenarios both with and without using a Secure Aggregator (SA). Additionally, the influence of impostor selections, with and without the incorporation of Generative Adversarial Network (GAN), is highlighted.
• Figure 5: The Equal Error Rate (EER) across 1000 devices is reported for the comparison between the individual and federated models in the unsupervised system.