Author(s): Marte Loge, Markus Duermuth, Lillian Rostad

Download: Paper (PDF)

Date: 12 Aug 2016

Document Type: Presentations

Associated Event: NDSS Symposium 2016

Abstract:

Android Unlock Patterns are one of the most widely used graphical password schemes. However, the scheme’s security is limited by users not choosing patterns uniformly but with a specific bias. In this work we take a closer look at this bias, in particular how personal traits influence the chosen patterns. We conducted a user study with 800 participants and demonstrate that certain factors such as age, gender, and experience in IT significantly influence the strength or length of the chosen patterns. This has implications both for how we can help users to select stronger patterns and for forensic applications.