Keystroke Biometric Authentication
on Smartphone Data
This project requires an understanding of biometrics and especially the keystroke biometric.
The literature review should include a review of gesture recognition since taps, swipes,
and other touch screen input movements are basically gestures.
This project will be a continuation of earlier work .
The work will focus on (and be restricted to) developing a robust keystroke biometric system for touch screen input on Android devices.
The keystroke dynamics for text input on handheld devices with their touch screens and soft keyboards will have considerably different characteristics
in contrast to the dynamics on desktop/laptop machines with hard keyboards.
In fact, on Android devices it is possible to enter more than one character with a single screen touch gesture.
Therefore, this study will investigate the viability of the keystroke biometric on handheld devices.
- Android smartphones will be provided for data collection and the collected data will be added to the existing data.
- Your customer will then run experiments on the data
to provide information as to how well the system performs on data captured on handheld devices.
- If time permits, improve the features of the system to improve system performance.
For example, how can we capitalize on the unique aspects of the small device touch screens (small form factor, how text is entered, etc.)
to design new and useful pattern recognition features.
- Michael J. Coakley, John V. Monaco, and Charles C. Tappert,
Numeric-Passcode Keystroke Biometric Studies on Smartphones,
Proc. Research Day, CSIS, Pace University, May 2015.