|Research field||Advanced Community Information Systems (ACIS)|
Multimedia capturing of collaborative scientific discourses about movies
Requirements for collaborative multimedia knowledge management systems can be summarized as follows. First, systems should emphasize the need to accept the full richness of media usage (1). Second, we want to capture any kind of disturbance (2) within all given media to promote learning. Third, flexible and adaptable transcription [JaSt02] techniques (3) for media usage are needed together with underlying formalisms (4). Thus, scientific discourses have to be carefully designed as communities of practice [Weng98]. They need the linkage of knowledge to action, with the implication of seamless de-contextualization from action and re-contextualization into action (5). Finally, interfaces to metadata description standards are necessary (6) that make an integration of media and models no longer an exotic adventure.
Figure 1: MECCA: Video screening platform for virtual communities
Approaches to capture semantics of multimedia artifacts
Semantics of digital multimedia materials are very hard to capture either manual or automatic. The implicit semantics of a multimedia artifact is the not necessarily shared set of interpretative acts of community members within a community of practice, which emerges by discursive assignment of semantics. The explicit semantics of a multimedia artifact is the set of terms created or linked in the practice of scientific discourses, which forms the multimedia ontology of the discourse (1). The development of a common practice integrates the negotiation of meaning between the members as well as the mutual engagement in joint enterprises and a shared repertoire of activities, symbols, and multimedia artifacts (5). By focusing on discursive knowledge creation processes with semantically enriched multimedia, we extract a community terminology (4) to categorize multimedia artifacts. Multimedia knowledge management systems must allow users to modify (transcribe) versatile media (3) in an arbitrary and interchangeable manner (5). These systems are capable of detecting conflicts (2) in the explicit semantics of an individual and its community (4) in the ongoing discourse (1) captured by metadata standards serving as the vocabulary of transcribed multimedia artifacts (6). Related work can be found e.g. at [DSUG03], [GoVi02], [GrSr01].
Knowledge management can be accomplished by the use of ontologies, having their primary area of application in the field of knowledge engineering instead of multimedia. An ontology in that sense is a terminological abstraction of the real world. Ontology-based systems have been developed to structure content and support retrieval. A problem is to find an ontology that fits into all users’ interpretations and allows a generic transcription of multimedia artifacts.
In contrast to strict classification of knowledge in common ontology-based systems or languages, the practice of multimedia management is the more or less unclassified handling on flat file systems. To support retrieval facilities one can annotate artifacts with metadata. In multimedia management systems metadata are supported systematically and efficient. The metadata used for semantic enrichment of the original information is of greater importance than before [Kosc02]. But, community artifacts and individual artifacts mix up and adequate retrieval of documents becomes more difficult. The risk of inappropriate addressing of multimedia artifacts increases (cf. figure 2), because user and community specific explicit semantics are not kept separate.
Figure 2: Traditional capturing of multimedia
Both, pure / multimedia ontology and community vocabulary driven creation of explicit semantics need an integration of the community terminology.
We introduce an approach that combines both, multimedia and community ontologies, deploying a standardized metadata vocabulary. The metadata description standard “Multimedia Content Description Interface” ISO/IEC 15938 1‑8 (MPEG‑7) [ISO02] has advanced features to describe and manage multimedia artifacts as well as multimedia collections. Hence, we deploy MPEG‑7 for the capturing of semantics giving users an unbiased view on multimedia artifacts.
MECCA: A MPEG-7 based video screening platform for virtual communities of practice
MECCA represents a multi-dimensional multimedia screening and classification platform. It’s our second implementation following the Virtual Entrepreneurship Lab (VEL) [KHJM02]. The basic idea has been taken from the original project “Berliner sehen” developed at the MIT [Fend01]. MECCA is specially designed for a community of researchers and students within a collaborative research center. For the researchers it serves as a media classification and monitoring system in a distributed setting. By students it is used as an e-learning system for video screenings of previously annotated media to share information about them independent of rare screening meetings in the university.
The scientists of MECCA’s media classification community work in a distributed setting. They have diverse backgrounds of educations, e.g. cinematic science, history of art, graphical design and are on diverse levels of profession, i.e. full professors, research assistants, and students. Due to their educational and cultural background, the community members have already different interests and point of views.
Figure 3: Capturing semantic relations in an ontological graph representation
In 2002 MECCA has been introduced to our colleagues. After several in-group meetings they defined six main categories and 30 subordinate categories (changed to five main categories and 28 subordinate categories later on; cf. figure 3) to classify a core set of 46 movie samples. We stored the core ontology in a MPEG-7 compliant database management system (cf. figure 4). Since the creation of new knowledge is a discursive and multistage process, users commonly took screenings on already existing multimedia materials. The next steps were done by gradually annotating and (re-)classifying multimedia artifacts. These annotations were managed the database. MECCA checked ontology structures and metadata enriched content descriptions automatically for conflicts and similarities on the structural and on the content level. The system classified those components matching fully or partially as well as those showing divergence. By doing so, we gave members within the heterogeneous MECCA community an option to express their problems with MECCA’s overall terminology. This kept a continuous discourse going on and created new implicit semantics.
Figure 4: Capturing multimedia discourses by combining multimedia artifacts and ontologies
Figure 1 shows MECCA’s front end used for video screenings of cinematic science students based on the previously defined classification scheme. To the right there the categorization schema, which allows metadata‑mediated browsing by switching between digital video sequences. We used an Apache web server connecting an XML database, which contains the discursively externalized multimedia semantics in MPEG-7 artifacts. Key-frame thumbnails of videos vary related to the chosen category. Dragging its key-frame and dropping it in the middle play a video. Besides, all annotations concerning the media file are now accessible. By highlighting the appropriate categories depending on the context of the selected media the user is encouraged to investigate other aspects associated with the media artifact. Users can collaboratively maintain collections (bottom left). Collections can be seen as an approach to focus on a certain aspect: combining and customizing media artifacts and compounding this view on single media components with additional, issue related information.
In short, the introduction of MECCA in a heterogeneous community of researchers was very successful. Currently, we undertake a case study with cinematic science students to evaluate whether MECCA may substitute conventional video screenings at the university.
[DSUG03] Dönderler, M. E., Saykol, E., Ulusoy, Ö., Güdükbay, U.: "BilVideo: A Video Database Management System", IEEE Multimedia, Multimedia-At-Work, Vol. 10, No. 1, (2003) pp. 66-70.
[Fend01] Fendt, K.: Contextualizing content. In Knecht, M., v. Hammerstein, K. (eds.): Languages across the curriculum. National East Asian Language Ctr., Columbus, Oh (2001), pp.201-223.
[ISO02] ISO/IEC: Information technology — Multimedia content description interface — Part 8: Extraction and use of MPEG-7 descriptions. International Organization for Standardization (2002).
[GoVi02] Golshani, F., Vissicaro, P.: Design of a Multi-Ethnic Dance Information Repository. B. Grosky (Ed.): Proceedings of SOFSEM 2002 Workshop on Multimedia Semantics, Milovy, Czech Republic, November 27-28, (2002).
[GrSr01] Grosky, W. I., Sreenath, D. V.: Metadata-Mediated Browsing and Retrieval in Semantically-Rich Cultural Image Collections. Proceedings of the 2001 Tokyo Symposium for Digital Silk Roads, Tokyo, Japan (December 2001).
[JaSt02] Jäger, L., Stanitzek, G. (eds.): Transkribieren - Medien/Lektüre. Wilhelm Fink Verlag, Munich (2002), in German.
[KHJM02] Klamma, R., Hollender, E., Jarke, M., Moog, P., Wulf, V.: Vigils in a Wilderness of Knowledge: Metadata in Learning Environments, Proceedings of IEEE International Conference on Advanced Learning Technologies (ICALT 2002), Kazan, Russia, September 9-12, 2002, IEEE Learning Technology Task Force, (2002) pp. 519-524.
[Kosc02] Kosch, H.: MPEG-7 and Multimedia Database Systems. SIGMOD Records, ACM Press, Vol. 31 (June 2000), pp. 34-39.
[Weng98] Wenger, E.: Communities of Practice – Learning, Meaning, and Identity. Cambridge University Press, Cambridge, UK (1998).
- Monika Pienkos
- Dominik Renzel
- Petra Löffler
- Thomas Meyer
- Joanna Barck
- Wolfgang Beilenhoff
- Sabine Hänsgen
- Prof. Dr. Matthias Jarke
- PD Dr. Ralf Klamma, AOR
- Dr. Marc Spaniol
- Marcus Stauff
- Thomas Waitz
Transkriptivität -- Operative Medientheorien als Grundlage von Informationssystemen in den Kulturwissenschaften
Published in H. Bublitz, R. Marek, C.L. Steinmann, H. Winkler (Hrsg.): Automatismen, pp. 299-314. Wilhelm Fink Verlag 2010. ISBD 978-3-7705-4987-0