Metrological Aspects in Inteligent Eye-tracking Systems

Eye-trackers are tools for measuring and recording the movement of the eyeballs. The advancement of technology implies their accuracy increase, which broadens their use in intelligent systems from supporting work, learning to medical diagnostics. The success of using the eye-tracker depends on selecting a device with appropriate metrological parameters. Meanwhile, no objective, standardized measurement methods give repeatable and reproducible test results. Therefore, comparing the equipment available on the market is difficult. To solve this problem, numerous research centres work on appropriate metrics resistant to the technical parameters of the instrument. Another issue is research into hardware properties with human observers. The individual variability of the participants unnecessarily increases the variance of the results. This article discusses both problems and presents a solution that allows to eliminate the human factor from the procedure for determining the metrological properties of the eye-tracker.