Face recognition is an important and active area of R+D whose interest has increased in recent years because of theoretical and application-driven motivations. From simple mobile phones unlocking mechanisms to complex secure biometric access control systems, face recognition is getting more and more attention. Due to the sensitivity of the involved biometric signals, privacy has shown to be a serious concern when working with digital imagery, especially for those systems that must process, recognize or classify face images (visual privacy).
Other approaches either obfuscate or partially encrypt the data, but they perform the recognition in the clear, and the recognition server has partial or total access to the sensitive user biometrics. By using advanced encryption techniques for the inputs, optimized for the encrypted processing of biometric data, CloudSEEP fully protects all the involved signals, both the user fresh biometrics (face images and biometric features extracted from those images) and the database templates. Therefore, CloudSEEP enables a secured outsourced biometric recognition that leaks no information at all to the recognition server, that works only with encrypted data.