DIMENSIONAL AFFECT AND EXPRESSION IN NATURAL AND MEDIATED INTERACTION

There is a perceived controversy as to whether the cognitive representation of affect is better modelled using a dimensional or categorical theory. This paper first suggests that these views are, in fact, compatible. The paper then discusses this theme and related issues in reference to a commonly stated application domain of research on human affect and expression: human computer interaction (HCI). The novel suggestion here is that a more realistic framing of studies of human affect in expression with reference to HCI and, particularly HCHI (Human-Computer-Human Interaction) entails some re-formulation of the approach to the basic phenomena themselves. This theme is illustrated with several examples from several recent research projects. The human face is involved in various aspects of verbal and non-verbal communication: acts of speech, facial expressions, facial gestures and movements, to list several major examples. Cognitive science, psychology, neuroscience, as well as sub-fields of information processing sciences are currently aimed at the study of facially mediated communication. Some of this work is aimed at developing new techniques in computer graphics, animation, and computer vision. Probably the most widely cited motivation for this research is to provide a more natural basis for humans to interact with computer systems via personification of artificially intelligent agents. While there has been considerable research along these lines, to date relatively few human computer interface technologies have successfully employed face processing. An overview of the status of HCI applications of face processing is beyond the scope of the present article, however a review of the situation has recently been given by Bartneck and Lyons (2007). In the current article I will focus nearly exclusively on some aspects of facial expression processing by humans and machines and how these relate to current research in human-computer interaction