Show me how you move, and I tell you who you are

Closed Human-Computer Interaction Computer Science eXtended Artifical Intelligence Master Project Thesis Internship No Affiliation

Background

Your motions can tell a lot about you. This can be very helpful, as we can use your motion profile for novel authentication systems to potentially replace, or at least augment, existing password-based solutions. However, it can also be very dangerous, as we can infer a lot of attributes, such as age, gender etc. from motion data as well. Our research focusses on shedding light on the bright and dark sides of drawing information from user motions with deep learning.

Within the scope of this relatively broad research topic you are free to focus on selected aspects of our research, including (but not limited to) the following:

• explore existing datasets, such as “Who is Alyx?” and the [“Berkeley Open Extended Reality Recording Dataset 2023”](https://rdi.berkeley.edu/metaverse/boxrr-23/] and develop tools and data visualizations that help to better understand and grasp these large datasets
• design new studies to either collect new data from participants or evaluate novel authentication or privacy protection approaches
• experiment with state of the art deep learning systems to improve existing approaches or draw new conclusions
• investigate new ways to improve usability of VR applications with this new technology
• find new data preprocessing strategies to improve efficiency of existing systems
• discuss and evaluate implications of our research on a higher level, e.g. in the context of privacy

If you want to do some research on your own, have a look at these works to get started:

• Extensible Motion-based Identification of XR Users using Non-Specific Motion Data
• Inferring Private Personal Attributes of Virtual Reality Users from Head and Hand Motion Data

There is a lot more you can do in the course of an internship, HCI Project or master thesis. If you are interested, feel free to contact Christian Rack!

Tasks

The project will focus on the following tasks:

• Data Exploration
• Experimenting with state of the art deep learning systems
• Designing studies to evaluate novel authentication mechanisms

Prerequisites

• Interest in state of the art research
• Basic machine learning knowledge
• Some Python skills

Contact Persons at the University Würzburg