Multi-FedLinks: Highly reliable decentralized multibiometric federated learning links

In recent years, the recognition accuracies of deep learning-based biometric recognition methods, which rely on large amounts of biometric data for training, have significantly increased. However, in practical applications, biometric data are often distributed in small and fragmented amounts among various local clients. Implementing distributed biometric recognition is therefore greatly important. Most existing distributed biometric methods are implemented by federated learning, and these methods suffer from two problems. (1) The current methods are overwhelmingly limited to addressing distributed single-biometric recognition problems and are not applicable to distributed multibiometric recognition. (2) The conversion from traditional local learning to distributed learning with multiterminal cooperation poses a series of security hazards that have not been addressed. To address these issues, a decentralized multibiometric federated learning links (Multi-FedLinks) model for distributed multibiometric recognition is proposed in this paper. The model consists of multiple FedLink structures, which are resistant to Byzantine attacks. Collaboration among the multiple FedLink structures is implemented with a third-party server to achieve multibiometric federated learning. Experimental results on the NUPT-FPV dataset demonstrate that the superiority of Multi-FedLinks model. Our code can be found in https://github.com/HYMu99/Multi-FedLinks.