Object recognition system guided by gaze of the user with a wearable eye tracker

Abstract

Existing approaches for object recognition typically rely on images captured on an ordinary digital camera and therefore the recognition task sometimes becomes difficult when the image contains other objects (cluttered) and the object of interest is not clearly indicated. In this work, we integrate a wearable eye tracker into the object recognition system in order to recognize which object the user is paying attention to in the scene camera. To demonstrate the usability of such a gaze based object recognition interface, we developed a prototypical application named Museum Guide 2.0 which can be used in a museum as a mechanical guide for visitors. © 2011 Springer-Verlag.