Gaze dependant prefetching of web content to increase speed and comfort of web browsing

Abstract

Video-oculography gaze tracking permits the monitoring of a computer user׳s point of regard (PoR) on the screen by tracking with a video camera the centre of the pupil and patterns of infrared reflections on the cornea. Gaze tracking has been used as an alternative to traditional pointing mechanisms during human computer interaction. Gaze direction provides also an implicit but imprecise hint about intention. This work uses gaze as a source of information about user׳s intention. We explore how to benefit from behind the scenes gaze tracking in a web browsing scenario by employing the user׳s PoR to prefetch URLs of potential future interest. Our working hypothesis is that before clicking on a URL of interest, the user would have previously fixated at the URL and/or the surrounding area. Hence, using gaze data to prefetch URLs can speed up web navigation and increase user comfort. The results of our experiments show that browsing the web using prefetching of web content based on gaze activity is in fact faster than traditional browsing. The caveat of our approach is that the network traffic volume increases as a result of the prefetching activity occurring in the browser׳s background. We have also carried out additional experiments to find out the optimal trade-off between speed gain and network traffic increase in a prefetching scenario. Leveraging upon the positive results of our experiments, we postulate the idea of an attentive computer system that monitors users gaze and takes preventive or assertive action in terms of what the user is likely to do next based on the direction of gaze. To our knowledge, this is the first in-depth study on the concept of prefetching content of potential future interest in advance using the information gathered by monitoring a user׳s gaze.