DBAC-DSR-BT: A secure and reliable deep speech recognition based-distributed biometric access control scheme over blockchain technology

Oussama Mounnan, Larbi Boubchir, Otman Manad, Abdelkrim El Mouatasim, Boubaker Daachi

Abstract: Highlights•A novel approach based on blockchain technology for biometric access control using speech recognition, allowing preserving the security and privacy of all system processes.•A modular, scalable and granular access control scheme that makes use of maintenance and upgrades more straightforward, by dividing the processes into two parts, two different smart contracts.•A novel biometric access control based on speech recognition using deep learning, namely the AE-GAN model, that enables getting the relevant and robust speech features for authentication operation.•A novel granular biometric access control that tackles existing system issues such as biometric template protection, and unreliable access control and authentication procedure, by using homomorphic encryption.•Implementation of the proposed prototype to prove its operationality, functionality, performance in terms of robustness and accuracy and resistance to several attacks.