1 Overall status

QMUL:
QMUL has implemented the classifier module based on Biologically Inspired System (particle swarm optimization and ant colony optimization techniques).

KU:
KU has specified a platform independent SemFs architecture (collaboration with Max Volkel from FZI in Karlsruhe). We have implemented a small prototype for browsing the SmartWeb Dataset of Football images to show how SemFs can be used with arbitrary domain data, and we have done some scalability tests.

ITI:
ITI has worked on using semantic metadata at the user interface level for a specific archeological search engine (mostly outside from kspace for
the moment). More specifically, ITI has developed a user interface that corresponds to a hybrid search engine that combines content-based and ontology based queries in a hybrid way, thus being capable of returning to the user results that are both visually and semantically similar. The interface is accessible from: http://mkg.iti.gr/reach/search.html but it is in Greek.

DCU:
DCU has designed an user interface for the query formulation tool used in the TRECVid experiments. The tool has been developed, tested and evaluated with 
users. Also DCU developed and tested an ontology-based video annotation tool, leading to a poster/paper submission to SAMT 2006 following the internship of Emilie GARNAUD from Eurecom to DCU.

CWI:
CWI has continued our work on the /facet prototype, improving nagivation possibilities and performance for large datasets. A demo/poster has been
submitted to ISWC 2006. /facet has also entered the Semantic Web Challenge 2006 in which it is actually in the Top 5 finalists. 
For more information and online demo: http://slashfacet.semanticweb.org/

CWI has done some work on the foundation paper of the journal special issue on "Canonical Processes of Media Production". In this work, nine basic processes, based on an examination of existing multimedia systems, have been identified (premeditate, create, annotate, package, query, construct message, organise, publish, and distribute). These processes should not be viewed as prepackaged, ready to be implemented by a programmer. Our goal is rather to analyse existing systems to identify and generalize functionality they provide and, on the basis of the processes supported within the system, determine which outputs should be available from the system. In this sense, we hope that system creators will be open to providing the outputs we identify when the processes are supported within the system.

GU:
GU has developed a video retrieval system that supports both explicit and implicit relevance feedback. One objective of our work was to find out if a combination of explicit and explicit relevance feedback returns better retrieval results than a system using explicit feedback only. 
For this purpose, we implemented two different video retrieval systems under the conditions of the international TRECVid workshop. These are two basic systems using both textual and visual queries for retrieval. Users can trigger retrieval cycles, browse through returned keyframes which represent video shots, play and scroll through the actual video file. For query refinement, users can give relevance feedback. This is helpful especially for query expansion. 
Both implemented systems had the same interface. For testing the stated hypothesis, we ran simulated user studies using the 24 search Topics of the 2005
TRECVid and evaluated the results.
Another approach was to identify a model to weight existing feature categories of implicit relevance feedback. Useful categories are:
  - an initial click on a keyframe (C1)
  - playing duration of a video file (c2)
  - interaction with a video (C3)
For identifying the best weighting model, we ran four simulated user studies using different values for C1 – C3. Again, we used the 24 TRECVid search topics from 2005 for evaluation.

2 Issues and Deviations from the Plan
None

3 Deliverables and Milestones for the Reporting Period
None

4 Dissemination
 - A demo/poster of /facet has been submitted to ISWC 2006 (CWI).
 - /facet has also entered the Semantic Web Challenge 2006 in which it is actually in the Top 5 finalists (CWI).
 - A joint EURECOM/DCU poster entitled "Evaluation of a Video Annotation Tool Based on the LSCOM Ontology" by Emilie Garnaud and Alan F. Smeaton has been submitted to SAMT 2006

5 Meetings Attendance
CWI has presented the results of the NewsML work integrated in the Facetted Browser during the IPTC Annual General Meeting, in Vienna, 
http://www.iptc.org/pages/meetg_main.php (2-6 of July).

All partners working in WP5.3 attended the K-Space meeting in Graz on 28 and 29 August. Many tools and systems have been developped and presented. A table of contents for D5.3 has been proposed and presented gathering all these systems.

6 Conclusion
Everything is going as planned