| Home | Research | Interests | Publications | IMS Home |
This page contains notes and ideas about the OCAR annoration application. It starts with a use case scenario to describe some functionality of the annotation application.
two users walk through the city. Their are interested in a certain architectural style '...' and have selected relevant keywords. Then both active their systems.
During the walk buildings with meta information matching the keywords are highlighted. In addition virtual icons appear are visible in certain locations and represent also available information for this place.
A list of matching items is also available to the user. This way they can explicitly highlight certain information. The list will probably be filtered on some level such as building or address.
the users can activate their 'eye-cursors' and look over any buildings or other geometry. a yellow sphere will appear where their gaze is intersecting with the real or virtual objects. If they look at a building more icons signifying localized information appear. If they look directly at an icon the associated information will appear directly in their screen.
This associated information takes three forms: text, image (or movie) and 3D data. Text and image data will be presented in an head-stabilized overlay like a dedicated data window to provide optimal quality. 3D objects can either be shown in a head-stabilized examine mode or world referenced of this is of interest. 3D registered images (for example images of older facades) are treated as 3D objects.
The users can also see the 'eye-cursor' of the other users. In addition they can switch on visual guides that help them find the other users cursor. It should also be possible see the information that is triggered by the other user.
This mode can be used for an explicit tourist guide scenario. All tourists are only slaves to the master application and display whatever the master has selected.
They are also able to permanently highlight a certain icon. This icon will appear highlighted also for the other user. This can be used to signify a certain piece of information to the other user.
each user can set virtual markers at the location of their 'eye-cursor'. These markers can be made visible to the other users as well. some attributes (color, size, shape, text ?) can be edited to be able to distinguish the markers. The users can establish some adhoc conventions to attach some meaning to different attributes.
We can add a function to directly get a path to any object on the list of available items. Also this target functionality could be coupled to the collaborative navigation application to allow to simultaneously guide another user. This would give a guide the possibility to control the tour simply by navigating the list of interesting objects.
A group of tourists are following a guide on a tour through the city. The guide selects some keywords fitting the topic of the tour. She then selects the first/closest item on the list and a path to this item is displayed in her view. Using the collaborative navigation she also makes sure all members of her group get the same path information displayed.
Upon arrival at the first local, the individual information icons appear. The guide selects on after the other and explains to the group some interesting details. She can support her stories by the detailed images provided by the system about features of the building.
The guide is running an annotation app in master mode and the tourists all have only a slave instance of this application. Moreover the tourists systems are all running the usuall navigation and collaboration app. These two are used by the guide to navigate the tourists. The tourists are all in the over-the-shoulder mode of the guides annotation app and therefore see exactly what she is selecting and viewing. However they don't have eye-cursors or similar interactions for themselves.
based on the example above, we could also do an automated tour by having an application controlling the other ones with preselected states. This could probably done with the APRIL framework. As such I would rather leave that as a possible project to the VRLU next fall.