Navigation auf


Department of Computational Linguistics

Webcam eye-tracking for ML applications


Webcam-based eye-tracking technology has improved in recent years, but has not yet been thoroughly tested for reading experiments. The aim of this project is to extend the WebQAmGaze dataset by collecting a control set in the lab, representing the upper bound of achievable data quality from webcams. The data is then used to improve fixation detection, fixation correction and/or gaze prediction algorithms to ensure webcam gaze data can be used in machine learning applications.


  • Python programming