Face detection and attentional frames for visually mediated interaction

Introduces a set of effective computational techniques for understanding the visual aspects of human interaction, which could be used, for example, in videoconferencing applications. First, we present methods for face detection and the capture of attentional frames to focus the processing for visually mediated interaction. Second, we present methods for recognising the various gesture phases that can be used to control the camera systems in the integrated system. Finally, we discuss how these techniques can be extended to "virtual groups" of multiple people interacting at multiple sites.