Human–Robot Interaction Through Robust Gaze Following

In this paper, a probabilistic solution for gaze following in the context of joint attention will be presented. Gaze following, in the sense of continuously measuring (with a greater or a lesser degree of anticipation) the head pose and gaze direction of an interlocutor so as to determine his/her focus of attention, is important in several important areas of computer vision applications, such as the development of nonintrusive gaze-tracking equipment for psychophysical experiments in Neuroscience, specialized telecommunication devices, Human–Computer Interfaces (HCI) and artificial cognitive systems for Human–Robot Interaction (HRI). We have developed a probabilistic solution that inherently deals with sensor models uncertainties and incomplete data. This solution comprises a hierarchical formulation of a set of detection classifiers that loosely follows how geometrical cues provided by facial features are used by the human perceptual system for gaze estimation. A quantitative analysis of the proposed architectures performance was undertaken through a set of experimental sessions. In these sessions, temporal sequences of moving human agents fixating a well-known point in space were grabbed by the stereovision setup of a robotic perception system, and then processed by the framework.