CHI 97 Electronic Publications: Organizational Overviews
Hypermedia Research at C&C Research Labs, NEC USA
Yoshinori Hara
C&C Research Laboratories
NEC USA, Inc.
110 Rio Robles
San Jose, CA 95134, USA
E-mail: hara@ccrl.sj.nec.com
Kojiro Watanabe
C&C Research Laboratories
NEC USA, Inc.
4 Independence Way
Princeton, NJ 08540, USA
E-mail: kojiro@ccrl.nj.nec.com
Abstract
In this organization overview, we explain our approach to next generation hypermedia environment. We discuss the importance of Computer-Human Interaction and ways to provide high level authoring and augmentation capabilities as key components for the environment. CHI will contribute to integrating hypermedia functions with media contents for improving information accessibility. Implemented systems such as multimedia stroryboarding tools demonstrate their advanced hypermedia capabilities.
Keywords:
hypermedia, storyboard, information visualization, media search, media query, information circulation
Introduction
A division of NEC USA, Computers & Communications Research Laboratories (CCRL) operates offices in Princeton, New Jersey and San Jose, California. The mission of CCRL is to identify essential technologies on the critical path toward building the 21st Century Information Society, to innovate and prove the effectiveness of new computer and communication technology. The number of researchers at CCRL are approximately 50. CCRL at the San Jose Office was established in August, 1995, focusing on research and development of next generation hypermedia and Internet technologies. The goal of the CCRL-San Jose research office is to provide significant hypermedia/multimedia authoring & augmentation software technologies, in particular, on the Internet.
© 1997 Copyright on this material is held by the authors.
CHI and Hypermedia
Clicking and advancing the "look & feel" interface are not the only Computer-Human Interaction issues for hypermedia. We consider CHI not only as an input/output method but also as a way of refining multimedia data. By refining the data and its organization in media, users gain useful and accurate information from the structure.
Figure 1: Well-organized Hypermedia through CHI.
Figure 1 shows a process of organizing well-structured hypermedia by applying CHI as a refinement tool. A person can author his/her own idea to put it into the hypermedia interactively. Then, he or she can augment another idea from the well-organized data. In some cases and at some levels, computers can automatically generate the structure and take advantage of its utility by applying media recognition and media understanding technologies. In order to coordinate these processes appropriately, CHI plays an important role in refining the hypermedia information.
Figure 2: Authoring &. Augmentattion on Hypermedia.
Figure 2 shows more detailed authoring and augmentation
processes on hypermedia/multimedia information. Media contents such as image,
video, and sound, are originally unstructured, and it is hard to utilize
information from them. By applying these processes, useful structures can be
implemented. The structure creation process can be divided into two categories: inter-media processing and intra-media processing. The difference
between the two comes from whether the relationship is among the objects
(scenes, whole images, etc.) or within the objects. Generally, we can specify
hypermedia functional capabilities (navigation, indexing, etc.) according to the
level of organization structure and/or the operation phase. In some cases, an
automatic generating approach may be applied.
Research Projects
At CCRL-San Jose, we have been conducting research and development on the following research projects which correspond to key functional capabilities as shown in Fig. 2. They are multimedia scenario writing support (electronic storyboarding), information visualization for the WWW, and content-based multimedia search and query, any of which can be combined as integrated systems.
(1) Multimedia Scenario Writing Support [1]
Scenario writing support is one of the key issues for inter-media organization.
We have been developing an authoring system called Anecdote in order to
support the pre-production process as well as the whole development process of
multimedia applications[1]. Anecdote employs the concept of
surrogate media and surrogate scene. Surrogate media is a type of
media, such as sketches, to represent the content of the media data to be
created. A surrogate scene can be a design sketch of a screen representing a
segment of the application structure. Both can be used to simulate the execution
of the scenario and can serve as the specification of the data to be created. By
substituting all surrogate objects with actual media contents, the application
becomes complete.
(2) Information Visualization for the WWW [2]
At present, there is no effective method of browsing through a Web locality.
Searching is appropriate when the user has a well-defined understanding of what
information is needed. However, in many cases, the user is not certain of
exactly what information is desired and needs to learn more about the content of
the Web locality. In this case, browsing is an ideal navigational strategy. One
of the most effective tools for browsing is overview diagrams. These diagrams
present the structure of the underlying information space and let users see
where they are, what other information is available and how to access it.
However, for the World-Wide Web, the overview diagrams represent large complex
network structures. They are generally shown as 2D or 3D graphs and
comprehending such large complex graphs is extremely difficult. Therefore, our
research is concerned with finding effective substructures from a complex
network structure of a web locality. We are using various techniques like access
log analysis and clustering for finding useful substructures. An example is the
focus+context views of WWW nodes[2]. This project is
also categorized as an inter-media processing, but more emphasis on augmentation
aspects.
(3) Content-based Multimedia Search and Query [3, 4]
Content-based multimedia search employs a structure as intra-media processing.
The search engine called COIR (Content-Oriented Information Retrieval)
[3] consists of two key components; media indexing
and media matching. These processes are done automatically using image
processing technologies. Our media indexing method is based on region division
approach, i.e., media objects are divided into several regions. Each region can
be assigned to a specific meaning. Then, media matching function calculates the
attribute values of each region to show the candidates in the order of
calculated similarity values. One of the advantages of this approach is to
provide object-level composite queries. That is, users can specify a query as
the combination of media features, media semantic information (keywords), and
the relationships among them. We have developed a visual query interface called
IFQ (In Frame Query)[4] and a content-based media
search engine for the WWW.
Future Plans
Since CCRL has sister organizations in Europe (Bonn, Berlin) and in Japan (Tokyo, Osaka), we are continuing to integrate these components to provide a new information circulation system, and evaluate useful applications on such a world-wide distributed environment.
As a next step, we will integrate all the necessary information gained by the augmentation or media understanding process. The stored information will, in turn, be reused for next authoring or media recognition process. This kind of authoring-augmentation loop is increasingly important as the multimedia objects become well-structured. CHI will contribute to integrating hypermedia functions with media contents for improving information accessibility.
References
- 1
-
K. Harada, et al. "Anecdote: A Multimedia Storyboarding System with Seamless Authoring Support," ACM Multimedia'96, pp.341-351, 1996.
- 2
-
S. Mukherjea, et al. "Focus+Context Views of World-Wide Web Nodes," to appear at ACM Hypertext'97, 1997
- 3
-
K. Hirata, et al. "Content-oriented Integration in Hypermedia System," ACM Hypertext'96, pp.11-22, 1996.
- 4
-
W. Li, et al. "IFQ: A Visual Query Interface for Object-based Image Retrieval," to appear at ACM CHI'97 (demonstration session), 1997.
CHI 97 Electronic Publications: Organizational Overviews
