Back to Job Hub

Thesis Subjects: Other

Hello Student,

Below you find Tobii student projects, suitable for a master thesis (if bachelor is not specified).

Do you find a project you are interested in? Please click the apply-button! Your application has to contain your CV, personal letter and transcript, otherwise it will not be considered. If possible, please also submit a link to your previous work (published papers, link to projects or coding etc). Do not forget to specify what project you are interested in! If you don't find a project that suits your education, you can apply through our open application ad.

Happy reading!


1. QA Tester (internship or bachelor thesis)

Hello there bug catcher!

We are currently looking for someone studying testing, test automation or QA to work with us on a thesis project or as an intern at Tobii Dynavox. Here you will get to work with first class eye-tracking technology, making sure that our hardware and software runs tip-top.

The project could involve:

  • Maintenance testing of our software (verify fixed bugs, regression testing)

  • Testing our latest new hardware

  • System verification of our software on the new hardware

You must be fluent in English as well as Swedish in both speaking and writing. Programming experience is a merit but not necessary.

Tobii Dynavox is developing both software and hardware to increase our consumer’s ability to communicate with their surroundings and to facilitate the use of computers and tablets. To learn more about Tobii, have a quick look at Tobii in 3 minutes.

2. Measure how gaze filters affect the outcome of wearable eye-tracking studies (master thesis)

Tobii Pro strives to provide the most advanced tools in the world for understanding human behavior. With the release of Tobii Pro Glasses 2, performing eye-tracking studies in real-world environments has become one of the main methods for extracting human attention data in many different fields of research.

One of the inherent problems in wearable eye-tracking is the cost of doing quantitative studies. Historically, any large-scale study has required several man-months of manual coding, i.e. mapping each gaze position from each recorded video to a static reference image. A common way of reducing the work-load is to apply fixation filters to the gaze data before starting the coding process. This typically reduces the coding time by a factor of 10.

The purpose of this Master’s Thesis project is to gain a deeper understanding of how Tobii Pro’s IV-T and attention filters work and to evaluate how they affect the end results of an eye-tracking study.

The project should use large data sets of coded gaze data, with and without filters, and extract all relevant types of attention metrics that are commonly used in the eye-tracking research field. This data should form the basis for a detailed statistical analysis of the error distributions related to each of the metrics.

Ideally, this will result in best-practice recommendations for how to use filters, together with a detailed analysis of how robust the different metrics are to changes in the data. Achieving a greater understanding of the statistical properties of wearable eye-tracking studies is of great value to Tobii Pro and will also benefit the eye-tracking research field as a whole.


We are looking for students with a strong background in mathematical statistics and scientific methodology. Basic programming knowledge (such as Matlab) is a requirement.

One of the key parts of this thesis is to understand general research methodologies and metrics in the field of wearable eye-tracking. The student will need to carefully design mathematical models, and to acquire necessary domain knowledge in order to achieve useful results.


3. Measure the impact of errors in automatically coded gaze data in wearable eye-tracking studies (master thesis)

Tobii Pro strives to provide the most advanced tools in the world for understanding human behavior. With the release of Tobii Pro Glasses 2, performing eye-tracking studies in real-world environments has become one of the main methods for extracting human attention data in many different fields of research.

One of the inherent problems in wearable eye-tracking is the cost of doing quantitative studies. Historically, any large-scale study has required several man-months of manual coding, i.e. mapping each gaze position from each recorded video to a static reference image. Tobii Pro’s Real-World Mapping tool uses advanced computer vision techniques to automate the coding process, saving many hours of manual work.

The purpose of this Master’s Thesis project is to gain a deep understanding of how errors in coded data affect the end results of an eye-tracking study. In order to do this, the different error types that can occur when running the Real-World Mapping tool have to be understood, and all the different elements of a quantitative eye-tracking study need to be modelled and simulated.

When a model of the system has been created, different distributions of errors and noise should be introduced to the coded gaze data and fed into the system. By measuring the impact of different error sources on the final results, we can determine the requirements on the quality of the automated mapping tools.

Understanding of the statistical properties of wearable eye-tracking studies will also benefit the eye-tracking research field as a whole.


We are looking for students with a strong background in mathematical statistics and scientific methodology. Basic programming knowledge (such as Matlab) is a requirement.

One of the key parts of this thesis is to understand general research methodologies and metrics in the field of wearable eye-tracking. The student will need to carefully design mathematical models, and to acquire necessary domain knowledge in order to achieve useful results.

Working at Tobii

At Tobii we love what we do - we work hard, play hard, and make every day exciting.

Learn more about what we do by viewing this short video Tobii in 3 minutes.