Keystroke Authentication With A Capacitive Display Using Different Mobile Devices

This study investigates keystroke dynamics as biometric authentication on different smartphones. We analysed different sensors in the smartphones which affect the error rates of the authentication. We also evaluate the effectiveness of different features based on the error rates. In addition, a framework is presented for using one device as a base model to authenticate the same person on other devices. We conduct with an experiment with three devices and three different keywords to assess how well different devices can be used (error rates smaller than 3.5 %) and suitable combinations of devices. Moreover, our experiment results showed that passwords spread over the whole keyboard have lower error rates.