Advancements and Applications of Federated Learning in Biometric Recognition

Zhengliang Lyu

2024

Abstract

The extensive use of biometric technologies has rendered the enhancement of model performance, while safeguarding user privacy, a significant concern. Federated learning, an emerging distributed machine learning technique, enhances model generalization and accuracy while safeguarding user privacy through collaborative training across several devices. This paper reviews the application progress of federated learning in the field of biometrics, and discusses its advantages in improving model performance and protecting user privacy, as well as the challenges and future development directions. Specifically, this paper first introduces different types of biometrics in the past and their disadvantages, the basic concepts and advantages of federated learning. It subsequently conducts a detailed analysis of the application of federated learning in biometrics, encompassing data diversity, real-time processing, and privacy protection. Then, this paper discusses the challenges faced by federated learning in biometrics, such as data leakage risk, high computing resource demand and uneven data distribution, and proposes optimization strategies such as edge cloud collaborative computing and distributed computing optimization. Finally, this article anticipates future advancements in federated learning within biometrics. The significance of this paper is to introduce the broad prospects of federated learning in the field of biometrics, and provide an effective method to protect user privacy for future research in biometrics

Paper Citation

in Harvard Style

Lyu Z. (2024). Advancements and Applications of Federated Learning in Biometric Recognition. In Proceedings of the 2nd International Conference on Data Analysis and Machine Learning - Volume 1: DAML; ISBN 978-989-758-754-2, SciTePress, pages 526-531. DOI: 10.5220/0013527700004619

in Bibtex Style

@conference{daml24,
author={Zhengliang Lyu},
title={Advancements and Applications of Federated Learning in Biometric Recognition},
booktitle={Proceedings of the 2nd International Conference on Data Analysis and Machine Learning - Volume 1: DAML},
year={2024},
pages={526-531},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013527700004619},
isbn={978-989-758-754-2},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 2nd International Conference on Data Analysis and Machine Learning - Volume 1: DAML
TI - Advancements and Applications of Federated Learning in Biometric Recognition
SN - 978-989-758-754-2
AU - Lyu Z.
PY - 2024
SP - 526
EP - 531
DO - 10.5220/0013527700004619
PB - SciTePress