"Tell Me Who You Are and I Will Tell You Your Unlock Pattern"

Abstract

Graphical passwords, like the Android Pattern Lock, are a popular security mechanism for mobile devices. The mechanism was proposed as an alternative to text-based passwords, since psychology studies have recognized that the human brain have a superior memory for remembering and recalling visual information.

This thesis aims to explore the hypothesis that human characteristics influence users choice of graphical passwords. A collection of 3393 user-created patterns were analysed in order to examine the correlation between people s choice of pattern and their characteristics, like hand size, age, gender and handedness.

This thesis first gives a detailed summary of related research on graphical passwords. Then it shows how an online survey was used for collecting user-selected passwords and information about the respondents. Lastly, the thesis explains how the data was analysed in terms of length and visual complexity in order to gain further insight in users choice of passwords.

Although the data could not provide significant evidence to accept the hypothesis, the results show that password strength significantly varies between gender, age and IT experience. Additionally, analysis of all the collected patterns shows a significant bias towards the selection of pattern starting position.