Abstract
This paper explores the ways in which head gestures can be applied to the user interface. Four categories of gestural task are considered, Pointing, Continuous Control, Spatial Selection and Symbolic Selection. For each, the problem is examined in the abstract, focusing on human factors and an analysis of the task, then solutions are presented which take into consideration sensing constraints and computational efficiency. A hybrid pointer control algorithm is described that is better suited for facial pointing than either pure rate control or pure position control approaches. Variations of the algorithm are described for scrolling and selection tasks. The primary contribution is to address a full range of interactive head gestures using a consistent approach which focuses as much on user and task constraints as on sensing considerations.