Transparent authentication scheme with adaptive biometrie features for IoT networks

With the comprehensive evolution of information communication technologies on mobile sensing objects, versatile ubiquitous networks embedded with specific-purpose sensors and intelligent wearable devices have promptly been developed and deployed, called the Internet of Things (IoT). On account of the popularity of IoT, the security issues have been promptly focused due to potential threats from IoT architectures. In consideration of the heterogeneous network property of IoT, in this paper we propose an authentication system which applies machine learning techniques to extract user bio-features as authentication tokens and transparently performs continual or real-time entity verification in the back-ground without the user's notices.