Monitoring Response Quality During Campimetry Via Eye-Tracking

This is the authors' abstract. We don't add key points for this paper.

In a variety of use-cases, deriving information on user's fatigue is an important step for content adaptation. In this work, we investigate which eye tracking related measures can predict the error rate (as a proxy of subject's fatigue)during a visual experiment. Data was collected during a 40 minutes campimetric task, where the user has to detect visual stimuli (i.e., dots) of different contrast. We found that eye-tracking measures can be used to train a machine learning model to predict the error rate of a user with an average correlation of 0.72±0.17. The results show that this method can be used to measure the user?s response quality.