BSc (15HP) Project: Development of a Software Pipeline for Eye Tracking in MRI-Compatible Experiments

Eye tracking is an important tool in neuroimaging research because it provides information about participants' visual attention, eye movements, and pupil responses during controlled experiments. When used together with MRI, eye-tracking data can complement brain-imaging measurements and provide additional information about participants' responses to visual stimuli.

The existing eye-tracking system available at USÖ and used in this project is designed for use in an MRI environment. However, the system provides the eye-tracking cameras' raw video feeds rather than ready-to-use measurements of pupil size and gaze position. This creates a need for software that can process the camera data and convert it into structured eye-tracking measurements.

Project goal

The goal of the project is to design and implement a software prototype that can acquire raw binocular eye-tracking camera data and transform it into time-stamped measurements of **pupil size and eye position**. The main emphasis is on software development, real-time data processing, computer vision/image processing, and system integration.

Main development tasks
1. Eye-tracking data acquisition
Develop software for receiving and processing the raw video streams from the eye-tracking system. The software should:

* acquire the camera streams;

* process the incoming frames in real time or near real time;

* maintain accurate timestamps for acquired frames;

* provide basic monitoring of the incoming data;

* handle missing or invalid frames in a robust manner.

2. Pupil detection and pupillometry
Implement an image-processing algorithm that identifies the pupil in each eye and estimates its size over time. The implementation should produce a continuous time series containing, at minimum:

* timestamp;

* estimated pupil size for the left eye;

* estimated pupil size for the right eye;

* quality/validity information indicating whether the measurement is reliable.

The students should investigate and evaluate appropriate image-processing techniques for detecting the pupil in the camera images and should consider practical issues such as noise, reflections, blinking, and partial occlusion.

3. Eye-movement and gaze tracking
Implement an algorithm for estimating the position of the pupil/eye in the camera image over time. The output should contain time-stamped coordinates for both eyes. Where feasible, the software should transform these coordinates into gaze coordinates relative to the visual display. The accuracy and robustness of the tracking under realistic experimental conditions should be evaluated.

Required knowledge and qualifications
The students should have:

* good programming skills in C++ and Python;

* good knowledge of DSA;

* an interest in real-time data processing and experimental systems.

Experience with **OpenCV, NumPy, scientific data processing, eye tracking, computer vision, or stimulus-presentation frameworks such as PsychoPy** is advantageous.

The project provides an opportunity to gain practical experience in developing a real-time computer vision system, working with experimental hardware, and producing research-quality data for subsequent scientific analysis.

Contacts: Paolo Forte (paolo.forte@oru.se) or Andrey Kiselev (andrey.kiselev@oru.se)