No Section...

VI.5 Sept.-Oct. 1999
Page: 17
Digital Citation

Research alerts

Jennifer Bruer

Joining Physical and Virtual Environments Through Mixed Reality Boundaries

Steve Benford, Boriana Koleva, Gail Reynard, and Chris Greenhalgh
School of Computer Science and Information Technology
The University of Nottingham
NG7 2RD, United Kingdom
{sdb, bnk, gtr, cmg}

Mixed reality research draws on techniques from augmented reality, augmented virtuality, tangible interfaces, and wearable computers to create new environments in which people can work integrally with physical and digital information. Our research is exploring a specific approach to mixed reality in which we join physical and virtual environments using common transparent boundaries.

A mixed reality boundary acts as a window between a physical and a virtual environment, leaving each distinct and adjacent and enabling the occupants of one to see into the other. One way of creating such a boundary is to project a fixed view of the virtual environment onto a wall in the physical environment while texture mapping a video view of the physical environment onto a virtual wall in the virtual environment and aligning the two views appropriately.

The Internet Foyer was an early demonstration of this technique. First, we created a virtual foyer, a 3-D visualization of an organization’s home pages located in a distributed virtual environment. Visitors to this visualization would see one another embodied as avatars, could talk to one another, and would also see representations of people accessing the pages through conventional web browsers. Second, this virtual foyer was joined to a physical reception area in the organization using a mixed reality boundary in order to create a common entry point into the organization’s physical and virtual infrastructure. The Internet Foyer was therefore a point at which the physical world and cyberspace came into direct and visible contact.

Figure 1 shows a view from the virtual foyer. The visualization can be seen in the center (green spheres are Web pages and blue arrows are links) with the video view of the remote foyer directly behind it. Another visitor’s avatar with a video face can be seen toward the bottom left. Figure 2 shows the corresponding view from the physical foyer.

In a second example, we used a mixed reality boundary to connect a virtual stage and a physical auditorium as part of a poetry performance. Our poet was immersed in a virtual world and appeared as an avatar on the virtual stage. The audience in the auditorium saw the virtual stage projected onto a wide screen. At the same time, the poet could look out of the stage at the audience through a video texture. Once it became clear to the audience that the poet could see them "through the screen," the poet was able to persuade them to participate in the performance, for example, standing up to answer questions, chanting, and gesturing. Figure 3 shows a moment from this performance in which the poet (his avatar can be seen in the center of the image) has just asked a member of the audience to stand up. Figure 4 shows the poet immersed in the virtual environment.

Just as there are different kinds of physical boundary with different properties (for example, doors, walls, and windows), so might there be different kinds of mixed reality boundary. In our most recent work we identified a set of properties through which mixed reality boundaries can be configured for a variety of uses. They include (1) permeability, the extent to which a boundary attenuates, amplifies, or transforms different kinds of information that pass across it, and(2) solidity, the extent to which a boundary can be traversed. Situational properties describe the spatial relationships between the boundary and the two connected spaces, including its location and alignment. Temporal properties are concerned with the lifetime and duration of the boundary. Finally, the metaproperties symmetry and representation concern the extent to which the other properties are configured to be the same on each side of the boundary and the manner in which they are represented to participants.

We are currently exploring how these properties can be realized, including the design of new virtual and physical boundary materials. Perhaps the greatest challenge here is creating nonsolid boundaries, that is, boundaries that allow a participant to pass from a virtual to a physical space and vice versa. Given that physical materialization and dematerialization are currently possible only in science fiction, we have chosen to focus on the more limited problem of creating boundaries that give participants and observers the illusion of passage from one side to the other. We are also exploring how multiple mixed reality boundaries can be used to join many physical and virtual spaces into a larger structure, a so-called tessellated mixed reality. These new developments are being brought together in the creation of a new performance work in collaboration with the performance company Blast Theory and the Center for Art and Media (Zentrum für Kunst und Medientechnologie—ZKM) in Germany as part of the European eRENA project. This work will exploit the novel physical properties of a number of "rain curtains," fine curtains of water into which views of a virtual world are projected and that can be physically traversed by performers and audience members. This work will see its first public performance in Autumn 1999.

Rotating Virtual Objects with Real Handles

Colin Ware and Jeff Rose
Faculty of Computer Science,
University of New Brunswick
P.O. Box 4400
Fredericton, E3B 5A3, New Brunswick, Canada

In 3-D graphics design systems, users must frequently position and orient virtual objects. It can be argued that the human factors problem of translational positioning of virtual objects has been essentially solved. A number of studies have shown that given a suitable device, a user can perform translational positioning rapidly and accurately either two- or three-dimensionally. Translational positioning can be accomplished despite the fact that the device held in the hand can be nothing like the object displayed and that the hand is typically held to the side of the computer, not on the screen where the object appears.

But similar studies have shown that virtual objects are rotated far more slowly—more than 10 seconds—than real objects, which can be rotated in one or two seconds. This paper describes a set of experimental studies designed to discover which of the differences between real object rotations and virtual object rotations were important in explaining the discrepancy, with a view to improving interfaces that require 3-D object positioning.

We used virtual "handle" objects, generated using computer graphics. The objects closely resembled the physical handles held by participants, as shown in Figure 1. By using a mirror reflecting the monitor screen, we were able to place the user’s hand holding the handle in the same place as the virtual image. Therefore, we could introduce various discrepancies between what was seen and what was felt, in order to find out which factors mattered most in influencing the speed of rotations. All viewing was with stereoscopic glasses.

We report the details of four experiments, the highlights of which are as follows. Our results showed that given the appropriate input and viewing conditions, users can rotate virtual objects almost as rapidly and accurately as they can rotate real objects. We showed that having the hand in the virtual work space is important. Subjects performed 35 percent more quickly when the real and virtual handle objects were in the same space, compared with displacing the handle manipulator laterally, as in most interfaces. When we used a sphere instead of a handle as an input object, performance was not reduced, despite the mismatch between what was seen and what was felt. However, a sphere is rather a special shape and thus it would be unwise to generalize to other kinds of shape mismatch.

Constructing, Organizing, and Visualizing Collections of Topically Related Web Resources

Loren Terveen, AT&T Labs - Research
Will Hill, AT&T Labs - Research
Brian Amento, AT&T Labs - Research and Department of Computer Science, Virginia Tech

Finding information is a common and fundamental task for users of the Internet. The tools commonly used for this task are search engines such as AltaVista and directories such as Yahoo. Users type in queries using keywords or navigate through hierarchical directories of topics, eventually obtaining (usually quite large) collections of Web pages that match their interests.

What then? How do users make sense of the results? No one can wade through hundreds or thousands of pages, and few people can even get through more than a handful. So how do you tell which are the best pages, or what kind of information each page contains, or whether all interesting pages are included? Furthermore, people who are truly interested in a topic usually want to boil it down to a manageable collection of perhaps 10 to 20 key sites that they will revisit and share with others. Current tools provide scant support for the process of exploring, comprehending, and organizing collections.

Our paper describes four innovations that address these user needs. First, we defined a higher-level unit of interaction, replacing the Web page with the site—a structured collection of pages, a multimedia document. A site is more appropriate for two reasons.

  1. A site usually contains a coherent body of content on a given topic (e.g., song lyrics, episode guides for a TV show, current weather conditions), divided into pages to ease navigation and download time. Thus, users want to know what’s available at a given site, not on a single page.
  2. Most hyperlinks to a site point to the "front door," or "splash," page, whereas most links from a site come from the site’s index page.

Second, we defined a new information structure, the clan graph, to represent collections of densely connected sites. The clan graph has a clear intuitive motivation based on concepts from social network analysis, social filtering, and cocitation analysis. A clan graph is defined in terms of a set of seed (example) sites specified by the user and is constructed by following hypertext links from the seeds. It is easy for users to specify seeds; they may get them from their bookmarks file, from an index page they found on the Web, or from a search engine. The clan graph construction algorithm is tolerant of "noise" in the seeds: a few off-topic seeds will not affect the quality of the graph.

Third, as our algorithm adds sites to the clan graph, it constructs site profiles. These profiles contain information about the amount and type of content contained on each site as well as the links between sites. The profile data help to inform user evaluation of the quality and function of a site.

Fourth, to enable users to comprehend and manage the information we extract, we have developed the auditorium visualization, which communicates important information, such as whether a site is structurally central or peripheral, whether a site is more of a content provider or an index, the internal structure of a site, and how sites link together. This interface supports users both in understanding a collection of sites and in organizing the collection (e.g., creating subcategories) to reflect their understanding.

This paper reports on detailed analysis and user studies that document the utility of this approach. We show that the clan graph construction algorithm tends to filter out irrelevant sites and discover additional relevant items. The auditorium visualization, augmented with drill-down capabilities to explore site profile data, helps users to find high-quality sites as well as sites that serve a particular function.


F1Figure 1. (left) Inside the virtual foyer

F2Figure 2. (right) The view from the physical foyer

F3Figure 3. (left): A poetry performance across a mixed reality boundary

F4Figure 4. (right): A poetry performance across a mixed reality boundary

UF1Figure. Participants in object rotation experiments saw virtual computer graphics images that matched in size, shape, and texture the wooden handles they held.

©1999 ACM  1072-5220/99/0900  $5.00

Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee.

The Digital Library is published by the Association for Computing Machinery. Copyright © 1999 ACM, Inc.


Post Comment

No Comments Found