Gaze-based real-time activity recognition for proactive interfaces

One active line of research on gaze-based interaction aims to predict user activities during interaction with computerized systems. All of the existing studies, however, are able to detect the performed activity only after the activity ends. Therefore, it is not possible to employ these systems in real-time proactive user interfaces. In this paper (1) an existing activity prediction system for pen-based mobile devices is modified for real-time activity prediction and (2) an alternative time-based activity prediction system is introduced. The results of our comprehensive experiments demonstrate that the newly developed system is more successful than the existing system with respect to real-time activity prediction.