Active Authentication Experiments Using Actual Application Usage Log
Gwonsang Ryu, Sohee Park, Daeseon Choi, Youngsam Kim, Seung-Hyun Kim, Soohyung Kim, Dowan Kim, Daeyong Kwon · 2018
Active authentication is a user authentication scheme that uses the user's behavioral characteristics and environmental information collected in the background, to reduce the number of explicit authentication requests. In this paper, we propose active authentication scheme, which performs Grade-Up and Grade-Extend after comparing mobile device user's confidence grade and level of application authentication. We collect application usage log, face confidence, Context of Interest (COI) familiarity, placement of mobile device and screen log from 22 participants over approximately 42.5 days. The proposed scheme classifies into 4 authentication levels applications that participants used depending on authentication method of the applications. Our experiments demonstrate that the proposed scheme is able to reduce the number of explicit authentication requests by 49%.