Combining Concept- with Content-based Multimedia Retrieval

The arrival of the XML standard opened new doors for structured document search. Common approach in XML retrieval is to directly exploit the documents structure. However this is likely to fail for two reasons. First of all, it neglects the rich multimedia character of documents on the Internet, where a wide variety of multimedia objects can be found such as text, images and streaming video. Secondly, using the document structure as the basis for searching the content of a document can easily lead to semantical misinterpretation of the document's content. This chapter discusses an approach for searching rich multimedia document collections, that tackles these two problems using a combination of conceptual search and content-based retrieval.