Keystroke-based Continuous Authentication while listening to Music on Your Smart-phone

摘要

Keystroke dynamics has long been proposed as a behavioral biometric for devices that collect input from keyboards, keypads, and virtual touch screen keypads. However, little work has been done on how user contexts impact this biometric. This work helps to fill this gap by examining how keystroke data collected in the presence of music would impact authentication with this biometric. To accomplish this, this work considers keystroke latencies (i.e. key hold, key press, and key interval) and uses two verifiers (i.e. relative measure and absolute measure) in determining the impact the presence of music has on smartphone keystroke data. The results show that, for the verifiers used, the presence of music has an impact on EERs and that some keystroke latencies (i.e. key hold and key interval) are more affected by the presence of music.