One of the most important bits of every search engine is the query interface. Complex interfaces may cause users to struggle in learning the handling. An example is the query language SQL. It is really powerful, but usually remains hidden to the common user. On the other hand the usage of current languages for Internet search engines is very simple and straightforward. Even beginners are able to find relevant documents. This paper presents a hybrid query language suitable for both image and text retrieval. It is very similar to those of a full text search engine but also includes some extensions required for content based image retrieval. The language is extensible to cover arbitrary feature vectors and handle fuzzy queries.

No abstract

This paper proposes a generic design for 3D model retrieval. It has been developed in cooperation with EADS, which deals with Computer Aided Design / Computer Aided Engineering (below CAD/CAE) data through the product life cycle. Sharing these models with users across the enterprise is a challenging task. CAD/CAE models may be hundreds of megabytes large and stored in proprietary formats. Browsing and previewing this data efficiently requires new tools.One way to leverage the collaboration in and outside of CAD domains is to offer through a repository an access to a neutral 3D data format. Also functionalities for semantic enrichment of 3D models, for the retrieval of context sensitive information and for 3D model retrieval based on 3D similarity search and CBIR techniques should be provided.

No abstract