Computer Vision for Interaction
Mathias Kölsch
Computer Science Department
University of California, Santa Barbara
Montag, 13.09.2004, 14 Uhr c.t., D6-135
Computer vision is becoming a viable user interface technology
for consumer-grade applications, opening tremendous
possibilities for novel devices and functionalities. Our
research has shown that reliable hand detection is feasible in
most lighting conditions and before most backgrounds. Also, we
built a fast tracking method that exploits optical flow and
color information to follow the location of a hand despite
rapid movements and concurrent finger articulations. Together
with appearance-based recognition of some key hand
configurations, this constitutes "HandVu," an out-of-the-box
system that recognizes hand gestures in real-time and acts as a
human interface device. It can complement or even replace
traditional input means, in particular for non-traditional
applications such as wearable computing and augmented
environments.
This talk will cover some of the underlying computer vision
methods, but also shed light on the potential of vision-based
interfaces for the future of human-computer interaction.
Bio
Mathias Kölsch recently completed his Ph.D. in Computer Science
at the University of California, Santa Barbara, where he was
advised by Professor Matthew Turk. Mathias' research centers
around computer vision, human-computer interaction, and
augmented reality. His goal is to widen the channel between
the human and the computer by building user interfaces that
bridge the "brain-computer barrier" as effectively as possible.
He has published in various conferences and has co-authored a
book chapter about perceptual interfaces. He has accepted a
faculty position at the Naval Postgraduate School in Monterey,
California.