User Identification based on Touch Dynamics

Touch interaction has quickly become the de-facto means of interacting with handheld devices due to its perceived attractiveness and low hardware cost. This study proposes a strategy for identifying users based on touch dynamics. Users' touch behavior is monitored and several unique features are extracted including left versus right hand dominance, one-handed versus bimanual operation, stroke size, stroke timing, symmetry, stroke speed and timing regularity. An experiment involving 20 users reveals that the strategy is successful in identifying users and their traits according to the touch dynamics. The results can be used for automatic user interface customization. However, more research is needed before touch characteristics can be applied to increasing the security of handheld touch-based devices.