Behauth is a client-side user authentication solution based on behavioral biometrics. Data collection, processing and authentication are performed solely on the user’s device, utilizing machine learning methods. The result of the authentication is securely sent to a remote party (e.g. a web application) that then takes appropriate actions.

Behauth analyzes a user’s movement patterns while freely interacting with a mobile device. An anomaly detection method is then employed to distinguish the device owner from impostors.

The solution also features an open-source library containing a comprehensive set of methods to process behavioral biometric data for future research to lay a common groundwork for further improvements of the biometric authentication accuracy.

The solution will be demonstrated as a smartphone application accessing a remote site requiring authentication.