| Summary: | In-home large displays such as TVs are becoming larger in size, and more
interactive in function. They start to be simultaneously used by multiple
people for various tasks in a dynamic setting. User interface issues such as
multiple users sharing the screen resources of the displays, and the usage
of multiple control devices have begun to emerge. We assume horizontally
laid out “personal interaction spaces” as the user interface for multiple users
to manage their screen real-estate. In this case, users often need to sign in
and out as well as have their personal spaces placed on the screen. Also,
the limited number and complex usage of conventional remote controllers
for TVs cannot satisfy the need of multiple interacting users.
In this thesis, we consider a
computer vision based system as a solution
to the emerging user interface issues. We built a vision system that tracks
the identities, positions and hand positions of people in front of a large
display to support our user studies of screen real-estate management and
multi-device management. We explore the usefulness of a vision system
through two user studies.
We designed the first study to compare the use of tracker-based mechanism versus manual ones for managing the display. Study Results suggest
that the tracking system is especially useful for simplifying the user sign
in/out process in conjunction with a manual method, and effective user-centric
placement of people’s interaction spaces.
As well, we designed a second study to explore whether contexts exist for
lower fidelity, gesture-based “remote controllers” for manipulating on-screen
objects. Study results show that gestural interfaces combined with high
fidelity devices such as a mobile phone in a group gaming scenario can be
useful for centralizing the control in the team and reducing errors. However,
gestural control is only suitable for simple, once-in-awhile interaction. === Applied Science, Faculty of === Electrical and Computer Engineering, Department of === Graduate
|
|---|
