| Popis: |
With increasing privacy concerns and security demands present within mobile devices, behavioral biometric solutions, such as touch based user recognition, have been researched as of recent. However, several vital contextual behavior factors (i.e., screen size, physical and application context) and how those effect user identification performance, remains unaddressed in previous studies. In this paper we first introduce a context-aware mobile user recognition method. Then a comparative experiment to evaluate the impacts of these factors in relation to user identification performance is presented. Experimental results have demonstrated that a user's touch screen usage behavior may be affected given different contextual behavior information. Furthermore, several interesting occurrences have been found in the results: 1) screen size of a smartphone device changes the way a user touches and holds the device. A larger screen size will provide more potential methods of interacting with the device and in effect, a higher user recognition accuracy as well; and 2) application context and physical activity context can aid in achieving higher accuracy for user recognition. |
