This installation of display, sensing and interaction technology is part of an on-going research collaboration between the College of Computing and Informatics and the Digital Arts Center of the College of Arts + Architecture.
Our goal is to use insights from anthropology and architecture to develop automated sensing techniques that are suited to public spaces and to develop strategies for the design of responsive virtual and physical environments.
Professor Chris Beorkrem- Digital Arts Center, SoA
Dr. Donna Lanclos, Anthropologist- University Library
Dr. Heather Lipford- Human Computer Interaction Lab, CCI
Professor Eric Sauda- Digital Arts Center, SoA
Dr. Richard Souvenir- Future Computig Lab, CCI
Public spaces, from the Vatican to Times Square have embraced the use of digital display systems as an important element of design. There is a significant body of research in social and ubiquitous computing that addresses how groups of people use computers or how computers are embedded in physical space. But there is no framework that deals with the design of combined virtual and physical spaces and the resulting interaction of large numbers of users as a social-computational system. Until such a framework exists, there will be no systematic method of designing or evaluating the use of computing in public places. As a result, many of these interactive public displays become merely expensive signage rather than enablers of interactive and engaging public spaces.
Our long term goal is to define a theoretical and practical framework for the design and evaluation of interactive computing in public spaces. We believe that for such spaces to better engage people and take advantage of virtual and physical capabilities, they should be adaptive to the people in them, incorporating not just a passive response to displays but rather interactivity in a particular place as a form of human centered computing.
The immediate objectives for this research are two-fold. First, is to develop automated sensing techniques that are particularly suited to such public spaces, using insights from anthropology and architecture to guide tracking and action recognition toward higher-level concepts of meaning, place and spatial affordance. Second, these insights will be used to develop new interaction paradigms that engage responsive virtual and physical environments and can afford multiple channels of communication between members of the public.
Our hypothesis is that camera-based sensing, when guided by a more sophisticated understanding of human behavior, is a general and inclusive method of capturing behavior in public spaces and that interaction systems based on human behavior will form a new model for HCI. Purely automated methods for video analysis are limited in their ability to extract meaning from object motion; thus there is a need to incorporate human understanding into a mixed-initiative framework to guide interface and architectural responses.
This research is an outgrowth of an on-going collaborative research group with researchers from human-computer interaction, computer vision, anthropology, and architecture. We have uncovered common issues and complementary methods and insights that will transform the use of interactive space. We believe that contributions from all of these fields will be essential to developing a convincing framework for computing in public spaces, and that our team is uniquely suited to address these issues. Our specific aims for this research are to:
- deploy camera networks in a real-world virtual and physical public spaces on a university campus;
- create a computer vision toolkit that captures, encodes, and visualizes human behavior;
- use the toolkit to determine meaningful sets of behaviors in the space, informed by and verified against standard ethnographic observations; and
- create design guidelines for responsive public environment that includes both HCI and architectural components.
This research is creative and original because it expands the scale of social computing to large public spaces and explicitly considers the interactions that occur at the intersection of virtual and physical space and how those spaces can be responsive. This approach will yield the following expected outcomes:
- advancements in vision algorithms for storing, analyzing, and presenting captured data;
- a flexible visual analytics toolkit aimed at designers and social scientists to help them develop and evaluate spaces; and
- strategies and design guidelines for responsive public spaces.