Abstract
The continuous proliferation of mobile and wearable smart devices, together with their increasing computational power and multitude of sensors, has given birth to innovative applications that enhance the world with virtual layers of processed information. In this paper, we present Talk2Me, an augmented reality social network framework that enables users to disseminate information in a distributed way and view others' information instantly. Talk2Me advertises users' messages, together with their face-signature, to every nearby device in a Device-to-Device fashion. When a user looks at nearby persons through her camera-enabled wearable devices (e.g., Google Glass), the framework automatically extracts the face-signature of the person of interest, compares it with the previously captured signatures, and presents the information shared by this person to the user. We design and implement Talk2Me to be lightweight, given that it runs on mobile devices with limited power. We analyze different content dissemination strategies to find the best protocol that yields reliable and fast information-spreading, while reducing the number of packets and containing the energy consumption on the devices. We design a novel face recognition algorithm for this specific scenario with a small number of face features and limited computing capability. Evaluation results of the prototype with real users and extensive simulations validate the performance and usability of our design, showing the potentials of the augmented reality social network framework in real-world scenarios.