The pervasive virtual world is becoming an important

By admin on October 27th, 2007

The pervasive virtual world is becoming an important part
of our lives. Unfortunately, we are only able to look at
information in books or on computer monitors. Computers,
projectors and video cameras have become sufficiently
affordable to allow for vision recognition and projection on
all the available surfaces of a space. These systems can
improve productivity and reduce the need for books and
portable computers, as well as allowing for more free space.
All five of the interfaces discussed were presented at the
Things That Think consortium meeting at the MIT Media
Laboratory in October 2003.
These interfaces stand to be considerably improved.
Information Table currently has only two modes:
countertop and dining table. Considering the numerous
functions possible on a work surface in a kitchen, the
system can only gain richness by being conceived for
additional uses.

Spatial Definition Some areas in the kitchen can

By admin on October 26th, 2007

Spatial Definition

Some areas in the kitchen can be dangerous to newcomers.
It would be helpful if certain conditions (such as a hot stove
or a knife on a table) could be easily denoted and
communicated. When someone enters these zones,
some automated multimedia reminder would make one
aware of unseen dangers. We have explored two ways for
such zones to be easily denoted and stored in a computer: a
kitchen design GUI and a gesture language.

The kitchen design GUI consists of plan of the kitchen
projected on a countertop with a drawing interface that
allows users to denote areas as dangerous. Users can also
annotate an area such as the table to show the menu when
users sit around it.
Gesture recognition techniques are widely used as
intuitive input interfaces. Our gesture language allows users
to specify a reminder zone directly on the space itself so
that anyone entering that zone triggers an audiovisual
reminder. Video cameras dispersed through the space
recognize a basic gesture language and create virtual zones
to map the physical environment accordingly. To find the
position of a user s hand, we use image processing
techniques that track color and motion in Visual C++. The
hand acts as a spatial pen so long as certain recognizable
gestures are made in front of the webcams. A back-
propagation Neural Network trains pre-defined shapes
(circles, triangles, arrows, and crosses.) The network
recognizes simple gestures if the path of the hand follows
one of the predetermined shapes. The gesture language for
spatial definition denotes an audio reminder zone with a
circle, a danger zone with a triangle, and a visual reminder
zone with an arrow.
Once a zone is programmed by the gesture language, the
same webcams serve as motion detectors to detect when
someone enters the zone. Each part of the kitchen can be
programmed with reminders without interfering with the
tasks being performed.