Encrypted sensing for enhancing security of biometric authentication

Abstract

In order to reduce the risk of theft or leakage of personal information captured by biometric sensing, we have proposed novel biometric sensing techniques in which biometric image can be captured using optical encryption methods. These methods, called encrypted sensing, are based on double random phase encoding (DRPE) and compressed sensing (CS). Encrypted sensing based on DEPR is implemented through digital holography (DH). On the other hand, encrypted sensing system based on CS is composed of digital micromirror device for random intensity modulation and single pixel camera. Through experiments using fingerprint images and finger vein images, we confirm that biometric images can be obtained as they are encrypted, and they can be restored correctly from the encrypted images.