Keystroke Dynamics
We used keystroke dynamics as biometric data to identify our users.
What are Keystroke Dynamics?
Keystroke dynamics measure the patterns of typing.
When you type, you spend fractions of a second between each downstroke and
upstroke—when you press and release each key. The time between your downstroke
and upstroke is “dwell time;” the time between upstroke and downstroke is “flight time.”
By recording your flight and dwell times for each key on your keyboard, we can find your
keystroke patterns, which identify you.
Figure 1: Graphic of Keystrokes.
Authentication
Current password research has found that combining multple authentication methods provides increased security. This sort of two-factor authentication can provide security against an attack where the password is already known by an attacker.
The three categories of authentication are as follows: what you know (a password), what you own (a token), and what you are (biometric data). We focused on this third option, because biometric data is unique to every person and doesn't require you to memorize or keep anything in addition to your password. It is much easier to use biometric data and a password than many other two-factor systems. Examples of biometric data include fingerprints, iris, handwriting and typing style.
Typing style is the easiest biometric data to gather on a standard laptop, so we resolved to use keystroke dynamics to add a second factor of security to a password. Our project built and tested models to recognize user keystrokes on multiple datasets of user keystroke information.