In the aceMedia project, novel knowledge structures and appropriate tools have been developed, since ontology structures have been extended and enriched to include low-level visual features and descriptors (following the MPEG-7 standard). This was done in order to support knowledge-aware multimedia analysis, media-aware semantic inferencing and high-level semantic reasoning, which will provide the means for automatic content annotation and generation of the ACE scalable metadata layer.

The developed knowledge infrastructure consists of several parts, namely the core ontology as a basis for all components of the knowledge infrastructure, multimedia-specific conceptualizations in form of the visual descriptor ontology (VDO) and the multimedia structure ontology and specific domain ontologies for selected personal and commercial domains. Additionally, a tool is needed for linking the prototypical instances of domain concepts from the domain ontology with corresponding instances of the VDO descriptors (i.e. linking with the multimedia/MPEG-7 descriptors).

Currently, there are many existing tools for ontology-based annotation. Most of them are document-centric, therefore they only support textual annotation of web pages and ontologies and there was no solution available so far, for the annotation of multimedia resources with low-level visual features. Thus, there was a need for the development of a tool that bridges and integrates the domain concepts and the low-level multimedia content description features.

In the aceMedia project, it has been decided to extend the CREAM (CREAting Metadata for the SemanticWeb) framework. OntoMat-Annotizer is the reference implementation of CREAM. It is Java-based and developed in the research field of ontology-based knowledge representation. There now exist two applications for the OntoMat-Annotizer framework: (i) it is used as an annotation tool for web pages and (ii) it acts as the basis of an ontology engineering environment. Moreover, OntoMat-Annotizer, provides a flexible plugin interface for further application extensions called OntoPlugin and offers the possibility to implement new components and extend the core functionality of OntoMat-Annotizer.

M-OntoMat-Annotizer Visual Editor and Media Viewer presents a graphical interface for loading and processing of visual content (images and videos), extraction of visual features and association with domain ontology concepts. Usually, the user needs to extract the features (multimedia descriptors) of a specific object inside the image/frame. For this reason, M-OntoMat-Annotizer lets the user draw a region of interest in the image and apply the multimedia descriptors extraction procedure only to the specific selected region. Alternatively, M-OntoMat-Annotizer supports automatic segmentation of the image; whenever an image is loaded it is automatically segmented into regions. User can then select a desired region or even merge two or more regions and proceed with the extraction.