Deep learning based cancelable biometric system

Rishabh Shukla, Harkeerat Kaur

Abstract: In this age of widespread cybercrime, simple measures like using a password or security token to verify your identity are no longer adequate. Over the past decade, biometric technologies have seen a meteoric rise in popularity as a dependable means of user authentication. This article discusses the introduction of a new technique dubbed "Cancelable Biometrics" for protecting the confidentiality of biometric data. A cancelable biometric solution (MLP) has been developed using deep learning techniques, which enables its use with multiple instances and provides erasable features. A cloud-based implementation of the proposed approach provides a revocable biometric authentication service accessible from any device. Biometric template security cancelability, low cost, High authentication accuracy, and quick response times are only few of the benefits of the described cancelable biometric system.