MSc Topic (30HP): Continuous Authentication in Virtual Reality Through Movement and Gaze Analysis

Background

VR systems typically authenticate users at the beginning of a session. However, if a legitimate user hands over the headset to another person, the new wearer may gain access to the authenticated session. Behavioral biometrics, such as head and hand movements and gaze patterns, may offer a way to detect such changes without interrupting the user.

The Project

The aim of this thesis is to investigate whether behavioral data can reveal that a different person has taken over an authenticated VR session, and how quickly such a change can be detected.

The student will develop a VR application using Unity and HTC Vive headsets with eye tracking. A group of participants will perform tasks in a virtual environment, including scenarios where one participant hands over the headset to another.

The collected movement and gaze data will be analyzed using suitable statistical and machine-learning methods. The evaluation will focus on detection accuracy, detection time, false alarms, and the contribution of different behavioral signals.

The project is scientific in nature and includes experimental design, data collection, analysis, and preparation of a scientific paper for submission to a relevant conference or journal.

Who We Are Looking For

We are looking for a Master's student in Computer Science, Computer Engineering, Cybersecurity, or a related field. Experience with Unity and C# is advantageous. Familiarity with machine learning, data analysis, VR, or cybersecurity is a plus.

The project offers an opportunity to combine VR development with experimental research at the intersection of cybersecurity and behavioral biometrics.

Practical Information

Scope: Master's thesis, 30 hp

Technology: Unity, C#, HTC Vive VR headsets with eye tracking

Method: VR application development, controlled user experiments, behavioral data analysis, and quantitative evaluation

Expected outcome: An experimental VR prototype, an empirical evaluation of behavior-based continuous authentication, and preparation of a scientific paper for submission.

Contact

Andrey Kiselev, andrey.kiselev@oru.se