Providing Dynamic Visual Information for Collaborative Tasks: Experiments With Automatic Camera Control

Abstract

One possibility presented by novel communication technologies is the ability for remotely located experts to provide guidance to others who are performing difficult technical tasks in the real world, such as medical procedures or engine repair. In these scenarios, video views and other visual information seem likely to be useful in the ongoing negotiation of shared understanding, or common ground, but actual results with experimental systems have been mixed. One difficulty in designing these systems is achieving a balance between close-up shots that allow for discussion of detail and wide shots that allow for orientation or establishing a mutual point of focus in a larger space. Achieving this balance can be difficult without disorienting or overloading task participants. In this article we present results from two experiments involving three automated camera control systems for remote repair tasks. Results show that a system providing both detailed and overview information was superior to systems providing only one or the other in terms of performance but that some participants preferred the detail-only system.

Notes

¹This is because of the nature of video. Multiple video cameras provide relative views of an essentially undefined global space. This lack of formal definition makes it hard to specify what is in view. For a more detailed discussion of these issues and possible emerging solutions, see Birnholtz et al. (2007).

Acknowledgments . We thank John Hancock, Xiang Cao, Clarissa Mak, Serena Kao, Tom Byuen, and members of the Dynamic Graphics Project Lab for their assistance with this research.

HCI Editorial Record . Received June 2, 2008. Revisions received April 30, 2009, and December 11, 2009. Accepted by Steve Whittaker. Final manuscript received April 12, 2010.