Use of Gaze Information for Deskwork Support System

Abstract

We aim to realize a deskwork support system that hands over a necessary object to a worker without any explicit instructions. In this study, we propose a method to predict the target object which the worker wants to get. Using gaze information as well as hand movements, the system can predict the target object appropriately at a very early stage A description of the experimental results demonstrates the usefulness of the method proposed here.