Personal Identification by Behavioral Features of Pinch-in and Pinch-out
Daiki Mukai, Ryuya Uda · 2018
Screen lock of smart phone can prevent malicious people from accessing the phone. However, it sometimes breaks by brute force attack or shoulder surfing so that security measures which resist unauthorized access and impersonation are required. Therefore, we propose a personal identification method with behavioral characteristics which is difficult to be copied by other people. In our proposal, behavioral characteristics of pinch-in and pinch-out while browsing are chosen for the identification and the characteristics are classified by deep learning. We have implemented the method as a system on a smart phone and we will evaluate how effective our method is against unauthorized access or impersonation.