Thesis Project Form

Commercial eye trackers have an accuracy of approximately 0.5 degrees. Even if this accuracy is reasonable in monocular tracking, the measurement error between the two eyes sums in binocular tracking, resulting in an unreliable estimate of the point of regard in the three-dimensional space.
The proposed thesis project will focus on developing hardware and software tools for binocular eye tracking, that would provide a robust and reliable estimate of the three-dimensional point of regard.

The objective of the thesis project is to analyze the criticisms of current technologies for binocular eye tracking and to develop possible solutions to increase accuracy, precision and usability of such technologies. The student will design and implement possible enhancements at hardware and/or software level and will test and validate their effectiveness against state-of-the-art technologies. These methodologies will be specifically aimed at increasing the accuracy and precision of the computation of the point of regard in the three-dimensional space, rather than on a flat two-dimensional computer screen. The obtained methodologies will be eventually used in case studies, like the use of the device in natural environments or for the analysis of vergence/version interactions.

The student will learn to employ different methodologies and instrumentations, including:
• Eye tracking
• Imaging technologies
• Computer vision
• Psychophysics