Titel of the Paper:

Metrosity - A realtime gestural tracking environment towards a personal interactive communication system

Authors:
Tony Brooks, Volker Krueger and Michel Guglielmi
Aalborg University

Extended Abstract:
Many people can be observed not being aware of their fellow commuters. A system
that encourages personal communication between people would give opportunities
for an improvement in life quality where the richness of human contact
could be realized. The abundance of online dating, chat rooms and other non
human communication forms point to the populous being aware of the need to
communicate. This paper presents the technical realization of an artistic interactive
vision project which is currently a ”work-in- progress” towards realization in
public spaces. The object of the vision project is to motivate people to interact
in a ‘human-to-human’ way.
Imagine, e.g., a scenario with two escalators descending, facing a big wall (see
Fig. 1). This is a scenario which often appears, e.g., in metro stations. Cameras
will be installed on the wall facing the escalators and the video stream will be
projected onto this wall. In that sense, the wall will become a big, artificial mirror
of the scenario facing it. If an individual enters the scene on the top of the escalators
and starts descending he/she sees ahead of him/her the whole environment
reflected on a wall, including the escalator he is standing on. By automatically
detecting the individual in the mirrored image, the individuals may be replaced
by avatars or, e.g., speech bubbles may be inserted into the mirror-image. In
that sense, a descending individual may notice that his reflection has an overlaid
image. This image is following him/her as the escalator descends. A similar
event may happen in the above scenario upon the other descending escalator.
The vision system can detect this event and can enter into the mirror-image
speech bobbles that may initiate some sort of communication between the two
individuals.

Reference:
1. A. Elgammal and L. Davis. Probabilistic framework for segmenting people under
occlusion. In ICCV, ICCV01, 2001.
2. D. Gavrila. The visual analysis of human movement: A survey. Computer Vision
and Image Understanding, 73:82–98, 1999.
3. M. Isard and A. Blake. Condensation – conditional density propagation for visual
tracking. Int. J. of Computer Vision, 1998.
4. D. Koller, J. Weber, and J. Malik. Robust multiple car tracking with occlusion
reasoning. In ECCV, ECCV94, 1994.
5. V. Krueger and S. Zhou. Exemplar-based face recognition from video. In Proc.
European Conf. on Computer Vision, Copenhagen, Denmark, June 27-31, 2002.
6. B. Li and R. Chellappa. Simultanious tracking and verification via sequential posterior
estimation. In Proc. IEEE Conf. on Computer Vision and Pattern Recognition,
Hilton Head Island, SC, June 13-15, 2000.