BOEMIE was an ambitious large-scale research effort that has advanced considerably the state of the art in multimedia content analysis. In order to make multimedia content like videos or images searchable the data must be meaningfully annotated. This is commonly done by humans, but it is a hard and expensive task. Using sophisticated algorithms to extract semantics from multimedia content, BOEMIE annotates content with semantics automatically and provides valuable knowledge for both, content providers and content consumers.
BOEMIE technology fuses knowledge that is automatically extracted by most of the popular media types (audio, video, images and text). In addition to making the content richer, the extracted knowledge is used to expand our understanding of the domain (for example athletics) and extract even more useful knowledge. This knowledge takes the form of ontologies and is represented in a machine-readable format. The synergetic process of extracting semantics from content and enriching the domain knowledge is the fundamental idea of bootstrapping that BOEMIE has pioneered.