The vision of the Internet-of-Things is rapidly changing the way we interact with everyday life objects, which can become not only data producers but also means for displaying information. This paper describes the design and implementation of such an object, a Smart Mirror that can be used to display personalised information to users in its proximity. An Android-based application is used to let users register to the Smart Mirror service, personalise the content they want to display and select whether such contents are public or private. To protect user's privacy and avoiding to disclose sensitive information to potential shoulder-surfers, the system is enriched with a BLE-based presence detection module following a device-free approach (i.e., without making use of any information from devices carried by users), which hides automatically private contents upon the detection of multiple people. The whole system is described in details and its performance are evaluated in realistic environments. In order to allow for reproducible research, the dataset used in this paper is made publicly available.
Design and Implementation of a Privacy-Aware Smart Mirror System
Longo, Edoardo;Redondi, AEC;Cesana, Matteo
2020-01-01
Abstract
The vision of the Internet-of-Things is rapidly changing the way we interact with everyday life objects, which can become not only data producers but also means for displaying information. This paper describes the design and implementation of such an object, a Smart Mirror that can be used to display personalised information to users in its proximity. An Android-based application is used to let users register to the Smart Mirror service, personalise the content they want to display and select whether such contents are public or private. To protect user's privacy and avoiding to disclose sensitive information to potential shoulder-surfers, the system is enriched with a BLE-based presence detection module following a device-free approach (i.e., without making use of any information from devices carried by users), which hides automatically private contents upon the detection of multiple people. The whole system is described in details and its performance are evaluated in realistic environments. In order to allow for reproducible research, the dataset used in this paper is made publicly available.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
