Face Transparent User Authentication Respecting Privacy

User authentication is an important issue to guarantee the security of information systems. For users, this task is in general inconvenient and boring. We propose in this paper a new transparent user authentication based on face recognition. During the use of a laptop or a smartphone, face features are computed, protected and sent to a trust party server that measures at any time the confidence of the user being the legitimate one. Authentication is transparent for the user while ensuring his/her security and privacy. We show through experiments with the AR face dataset the benefit of the proposed approach.