Mechanisms of intentional joint visual attention

Abstract

People communicate with others via intention. This is likewise true for the primitive behavior of joint visual attention: directing one’s attention to an object another person is looking at. However, the mechanism by which intention, a kind of internal state, causes that behavior is unclear. In this paper, we construct a simple computational model for examining these mechanisms, and investigate mechanisms for categorizing visual input and for recalling and comparing between these categories. In addition, we lay out some interaction experiments involving a human and a robot equipped with the constructed computational model, to serve as a platform for verifying the intentionality demonstrated by these mechanisms.