Proceedings of the 2nd Workshop on Adaptive Hypertext and Hypermedia
HYPERTEXT'98, Pittsburgh, USA, June 20-24, 1998

Automated Generation of Hypermedia Presentations from Pre-existing, Tagged Media Objects

Lynda Hardman, Lloyd Rutledge, and Dick Bulterman
INS, CWI, Kruislaan 413, 1098 SJ Amsterdam, The Netherlands
http://www.cwi.nl/~lynda/
Lynda.Hardman@cwi.nl
tel +31-20-592 4127
fax +31-20-592 4199

A standard reference model for the automated creation of multimedia presentations has been developed [BMFR98]. This is a high level framework for describing the division of processes and the communication among processes required for the automated creation of multimedia presentations.

The work is based on several operational projects for creating presentations with multiple media. The underlying document model assumed in the reference model lacked, however, a true time dimension and hypertext links. Also, it assumes sufficient knowledge of the domain and sufficient processing power to enable the constituent media items in the final presentation to be generated from scratch. Our approach is different in two respects. First, in the more complex underlying document model and, second, that the media items to be used in the presentation already exist. In addition, we are also investigating the utilisation of existing tools and standards for implementing some of the generation processes, as described in the SRM.

Our work on integrating a more complex document model is based on the Amsterdam Hypermedia Model (AHM), and has resulted in [HaWB98]. This describes how the different processes in [BFMR98] can be seen as ``filling in'' different parts of the AHM document model. These include the media items to be used, their temporal and spatial layout, style information and composition information. Relevant to this is the process of deciding on navigation structures among the various scenes of the presentation, in the line of HDM [GaPS93]. Work is currently being carried out in this area by a masters student.

We have also carried out our own system design for the generation of hypermedia presentations, based on the availability of tagged media items stored in a database [WBHT97]. This work is being continued with the implementation of a specialist tagging tool built on top of an existing image database. Video will also be incorporated. Two undergraduate students and a graduate student are working on the database and video aspects of this project.

We have recently started work on using style sheets, including structural transformations, for implementing the processes specified in the SRM. In particular as a vehicule for communication among the content, design, realization and presentation display layers. This has resulted in the submission of [RHOB98].

In broad terms, we see the process of generating a hypermedia presentation as follows.

A database contains a collection of media items. Each of these is tagged with one or more terms from a domain-specific thesaurus. The name of the source of the tagging is also included, since a single media item may be tagged with terms from multiple thesaurus.

As the end-user explores the information presented on the screen they are at some point in ``conceptual space''. This can be described by one or more terms from the domain-specific thesaurus. When the user requests further information, the system uses known information about the user (such as what has already been seen) and searches the accessible databases for relevant information. The search is based on the terms associated with the media items in the database(s).

Once a number of items have been found, they are grouped together into scenes which can be presented on the screen. Not all items can be presented at once, and choices have to be made on the temporal, spatial and linking relationships used to present the information in the best (for the end-user) possible way. Heuristics for grouping the information, and for temporal and spatial alignment can be expressed using a style sheet language such as DSSSL [ISO96].

Our work is in progress and we are approaching the problem from a number of different directions.

References

[BFMR98] M. Bordegoni, G. Faconti, M.T. Maybury, T. Rist, S. Ruggieri, P. Trahanias and M. Wilson (1998). A Standard Reference Model for Intelligent Multimedia Presentation Systems. Computer Standards and Interfaces.

[GaPS93] F. Garzotto, P. Paolini and D. Schwabe (1993). HDM -- A Model-Based Approach to Hypertext Application Design. TOIS 11(1) 1-26.

[HaWB98] L. Hardman, M. Worring and D.C.A. Bulterman (1998). Integrating the Amsterdam Hypermedia Model with the Standard Reference Model for Intelligent Presentation Systems. Computer Standards and Interfaces. http://www.cwi.nl/ftp/mmpapers/AHM-SRM.ps.gz

[ISO96] International Standards Organization. Document Style Semantics and Specification Language (DSSSL). ISO/IEC IS 10179:1996, 1996.

[RHOB98] L. Rutledge, L. Hardman, J. van Ossenbruggen and D.C.A. Bulterman (1998). Implementing Adaptability in the Standard Reference Model. Submitted to Multimedia Modelling 1998.

[WBHT97] M. Worring, C. van den Berg, L. Hardman and A. Tam (1997). System Design for Structured Hypermedia Generation. Lecture Notes in Computer Science 1306, Visual Information Systems, ed. C. Leung. http://www.cwi.nl/ftp/mmpapers/LNCS-visual.ps.gz